“Artificial Intelligence must be about more than our things. It must be about more than our machines. It must be a way to advance human behavior in complex human situations.But this will require wisdom-powered code. It will require imprinting AI’s genome with social intelligence for human interaction. It must begin right now.” — Phil Lawson (read more)
The other night, my nine-year-old daughter (who is, of course, the most tech-savvy person in the house), introduced me to a new Amazon Alexa skill.
“Alexa, start a conversation,” she said.
We were immediately drawn into an experience with new bot, or, as the technologists would say, “conversational user interface” (CUI). It was, we were told, the recent winner in an Amazon AI competition from the University of Washington.
At first, the experience was fun, but when we chose to explore a technology topic, the bot responded, “have you heard of Net Neutrality?” What we experienced thereafter was slightly discomforting.
The bot seemingly innocuously cited a number of articles that she “had read on the web” about the FCC, Ajit Pai, and the issue of net neutrality. But here’s the thing: All four articles she recommended had a distinct and clear anti-Ajit Pai bias.
Now, the topic of Net Neutrality is a heated one and many smart people make valid points on both sides, including Fred Wilson and Ben Thompson. That is how it should be.
But the experience of the Alexa CUI should give you pause, as it did me.
To someone with limited familiarity with the topic of net neutrality, the voice seemed soothing and the information unbiased. But if you have a familiarity with the topic, you might start to wonder, “wait … am I being manipulated on this topic by an Amazon-owned AI engine to help the company achieve its own policy objectives?”
The experience highlights some of the risks of the AI-powered future into which we are hurtling at warp speed.
If you are going to trust your decision-making to a centralized AI source, you need to have 100 percent confidence in:
The integrity and security of the data (are the inputs accurate and reliable, and can they be manipulated or stolen?)
The machine learning algorithms that inform the AI (are they prone to excessive error or bias, and can they be inspected?)
The AI’s interface (does it reliably represent the output of the AI and effectively capture new data?)
In a centralized, closed model of AI, you are asked to implicitly trust in each layer without knowing what is going on behind the curtains.
You see a man walking toward you on the street. He reminds you of someone from long ago. Such as a high school classmate, who belonged to the football team? Wasn’t a great player but you were fond of him then. You don’t recall him attending fifth, 10th and 20th reunions. He must have moved away and established his life there and cut off his ties to his friends here.
You look at his face and you really can’t tell if it’s Bob for sure. You had forgotten many of his key features and this man seems to have gained some weight.
The distance between the two of you is quickly closing and your mind is running at full speed trying to decide if it is Bob.
At this moment, you have a few choices. A decision tree will emerge and you will need to choose one of the available options.
In the logic diagram I show, there are some question that is influenced by the emotion. B2) “Nah, let’s forget it” and C) and D) are results of emotional decisions and have little to do with fact this may be Bob or not.
The human decision-making process is often influenced by emotion, which is often independent of fact.
You decision to drop the idea of meeting Bob after so many years is caused by shyness, laziness and/or avoiding some embarrassment in case this man is not Bob. The more you think about this decision-making process, less sure you’d become. After all, if you and Bob hadn’t spoken for 20 years, maybe we should leave the whole thing alone.
Thus, this is clearly the result of human intelligence working.
If this were artificial intelligence, chances are decisions B2, C and D wouldn’t happen. Machines today at their infantile stage of development do not know such emotional feeling as “too much trouble,” hesitation due to fear of failing (Bob says he isn’t Bob), or laziness and or “too complicated.” In some distant time, these complex feelings and deeds driven by the emotion would be realized, I hope. But, not now.
At this point of the state of art of AI, a machine would not hesitate once it makes a decision. That’s because it cannot hesitate. Hesitation is a complex emotional decision that a machine simply cannot perform.
There you see a huge crevice between the human intelligence and AI.
In fact, animals (remember we are also an animal) display complex emotional decisions daily. Now, are you getting some feeling about human intelligence and AI?
Shintaro “Sam” Asano was named by the Massachusetts Institute of Technology in 2011 as one of the 10 most influential inventors of the 20th century who improved our lives. He is a businessman and inventor in the field of electronics and mechanical systems who is credited as the inventor of the portable fax machine.
A few highlights from THE BUSINESS INSIDER INTERVIEW with Jaron
Butthat general principle — that we’re not treating people well enough with digital systems — still bothers me. I do still think that is very true. —
Well, this is maybe the greatest tragedy in the history of computing, and it goes like this: there was a well-intentioned, sweet movement in the ‘80s to try to make everything online free. And it started with free software and then it was free music, free news, and other free services.
But, at the same time, it’s not like people were clamoring for the government to do it or some sort of socialist solution. If you say, well, we want to have entrepreneurship and capitalism, but we also want it to be free, those two things are somewhat in conflict, and there’s only one way to bridge that gap, and it’s through the advertising model.
And advertising became the model of online information, which is kind of crazy. But here’s the problem: if you start out with advertising, if you start out by saying what I’m going to do is place an ad for a car or whatever, gradually, not because of any evil plan — just because they’re trying to make their algorithms work as well as possible and maximize their shareholders value and because computers are getting faster and faster and more effective algorithms —
what starts out as advertising morphs into behavior modification.
A second issue is that people who participate in a system of this time, since everything is free since it’s all being monetized, what reward can you get?Ultimately, this system creates assholes, because if being an asshole gets you attention,that’s exactly what you’re going to do. Because there’s a bias for negative emotions to work better in engagement, because the attention economy brings out the asshole in a lot of other people, the people who want to disrupt and destroy get a lot more efficiency for their spend than the people who might be trying to build up and preserve and improve.
Q: What do you think about programmers using consciously addicting techniques to keep people hooked to their products?
A: There’s a long and interesting history that goes back to the 19th century, with the science of Behaviorism that arose to study living things as though they were machines.
Behaviorists had this feeling that I think might be a little like this godlike feeling that overcomes some hackers these days, where they feel totally godlike as though they have the keys to everything and can control people
— I think our responsibility as engineersis to engineer as well as possible, and to engineer as well as possible, you have to treat the thing you’re engineering as a product.
You can’t respect it in a deified way.
It goes in the reverse. We’ve been talking about the behaviorist approach to people, and manipulating people with addictive loops as we currently do with online systems.
In this case, you’re treating people as objects.
It’s the flipside of treating machines as people, as AI does. They go together. Both of them are mistakes
Various computer scientists, researchers, lawyers and other techies have recently been attending bi-monthly meetings in Montreal to discuss life’s big questions — as they relate to our increasingly intelligent machines.
Should a computer give medical advice? Is it acceptable for the legal system to use algorithms in order to decide whether convicts get paroled? Can an artificial agent that spouts racial slurs be held culpable?
And perhaps most pressing for many people: Are Facebook and other social media applications capable of knowing when a user is depressed or suffering a manic episode — and are these people being targeted with online advertisements in order to exploit them at their most vulnerable?
Researchers such as Abhishek Gupta are trying to help Montreal lead the world in ensuring AI is developed responsibly.
“The spotlight of the world is on (Montreal),” said Gupta, an AI ethics researcher at McGill University who is also a software developer in cybersecurity at Ericsson.
His bi-monthly “AI ethics meet-up” brings together people from around the city who want to influence the way researchers are thinking about machine-learning.
“In the past two months we’ve had six new AI labs open in Montreal,” Gupta said. “It makes complete sense we would also be the ones who would help guide the discussion on how to do it ethically.”
In November, Gupta and Universite de Montreal researchers helped create the Montreal Declaration for a Responsible Development of Artificial Intelligence, which is a series of principles seeking to guide the evolution of AI in the city and across the planet.
Its principles are broken down into seven themes: well-being, autonomy, justice, privacy, knowledge, democracy and responsibility.
“How do we ensure that the benefits of AI are available to everyone?” Gupta asked his group. “What types of legal decisions can we delegate to AI?”
Doina Precup, a McGill University computer science professor and the Montreal head of DeepMind … said the global industryis starting to be preoccupied with the societal consequences of machine-learning, and Canadian values encourage the discussion.
“Montreal is a little ahead because we are in Canada,” Precup said. “Canada, compared to other parts of the world, has a different set of values that are more oriented towards ensuring everybody’s wellness. The background and culture of the country and the city matter a lot.”
As autonomous and intelligent systems become more pervasive, it is essential the designers and developers behind them stop to consider the ethical considerations of what they are unleashing.
That’s the view of the Institute of Electrical and Electronics Engineers (IEEE) which this week released for feedback its second Ethically Aligned Design document in an attempt
to ensure such systems “remain human-centric”.
“These systems have to behave in a way that is beneficial to people beyond reaching functional goals and addressing technical problems. This will allow for an elevated level of trust between people and technology that is needed for its fruitful, pervasive use in our daily lives,” the document states.
“Defining what exactly ‘right’ and ‘good’ are in a digital future is a question of great complexity that places us at the intersection of technology and ethics,”
“Throwing our hands up in air crying ‘it’s too hard’ while we sit back and watch technology careen us forward into a future that happens to us, rather than one we create, is hardly a viable option.
“This publication is a truly game-changing and promising first step in a direction – which has often felt long in coming – toward breaking the protective wall of specialisation that has allowed technologists to disassociate from the societal impacts of their technologies.”
“It will demand that future tech leaders begin to take responsibility for and think deeply about the non-technical impact on disempowered groups, on privacy and justice, on physical and mental health, right down to unpacking hidden biases and moral implications. It represents a positive step toward ensuring the technology we build as humans genuinely benefits us and our planet,” [University of Sydney software engineering Professor Rafael Calvo.]
“We believe explicitly aligning technology with ethical values will help advance innovation with these new tools while diminishing fear in the process” the IEEE said.
Jordi Ribas, left, and Kristina Behr, right, showcased Microsoft’s AI advances at an event Wednesday. Photo by Dan DeLong.
These days, people want more intelligent answers: Maybe they’d like to gather the pros and cons of a certain exercise plan or figure out whether the latest Marvel movie is worth seeing. They might even turn to their favorite search tool with only the vaguest of requests, such as, “I’m hungry.”
When people make requests like that, they don’t just want a list of websites. They might want a personalized answer,such as restaurant recommendations based on the city they are traveling in. Or they might want a variety of answers, so they can get different perspectives on a topic. They might even need help figuring out the right question to ask.
At a Microsoft event in San Francisco on Wednesday, Microsoft executives showcased a number of advances in its Bing search engine, Cortana intelligent assistant and Microsoft Office 365 productivity tools that use artificial intelligence to help people get more nuanced information and assist with more complex needs.
“AI has come a long way in the ability to find information, but making sense of that information is the real challenge,”said Kristina Behr, a partner design and planning program manager with Microsoft’s Artificial Intelligence and Research group.
Microsoft demonstrated some of the most recent AI-driven advances in intelligent search that are aimed at giving people richer, more useful information.
Another new, AI-driven advance in Bing is aimed at getting people multiple viewpoints on a search query that might be more subjective.
For example, if you ask Bing “is cholesterol bad,” you’ll see two different perspectives on that question.
That’s part of Microsoft’s effort to acknowledge that sometimes a question doesn’t have a clear black and white answer.
“As Bing, what we want to do is we want to provide the best results from the overall web. We want to be able to find the answers and the results that are the most comprehensive, the most relevant and the most trustworthy,” Ribas said.
“Often people are seeking answers that go beyond something that is a mathematical equation. We want to be able to frame those opinions and articulate them in a way that’s also balanced and objective.”
This article attempts to bring our readers to Kate’s brilliant Keynote speech at NIPS 2017. It talks about different forms of bias in Machine Learning systems and the ways to tackle such problems.
The rise of Machine Learning is every bit as far reaching as the rise of computing itself.
A vast new ecosystem of techniques and infrastructure are emerging in the field of machine learning and we are just beginning to learn their full capabilities. But with the exciting things that people can do, there are some really concerning problems arising.
Forms of bias, stereotyping and unfair determination are being found in machine vision systems, object recognition models, and in natural language processing and word embeddings. High profile news stories about bias have been on the rise, from women being less likely to be shown high paying jobs to gender bias and object recognition datasets like MS COCO, to racial disparities in education AI systems.
What is bias?
Bias is a skew that produces a type of harm.
Where does bias come from?
Commonly from Training data. It can be incomplete, biased or otherwise skewed. It can draw from non-representative samples that are wholly defined before use. Sometimes it is not obvious because it was constructed in a non-transparent way. In addition to human labeling, other ways that human biases and cultural assumptions can creep in ending up in exclusion or overrepresentation of subpopulation. Case in point: stop-and-frisk program data used as training data by an ML system. This dataset was biased due to systemic racial discrimination in policing.
Harms of allocation
Majority of the literature understand bias as harms of allocation. Allocative harm is when a system allocates or withholds certain groups, an opportunity or resource. It is an economically oriented view primarily. Eg: who gets a mortgage, loan etc.
Allocation is immediate, it is a time-bound moment of decision making. It is readily quantifiable. In other words, it raises questions of fairness and justice in discrete and specific transactions.
Harms of representation
It gets tricky when it comes to systems that represent society but don’t allocate resources. These are representational harms. When systems reinforce the subordination of certain groups along the lines of identity like race, class, gender etc.
It is a long-term process that affects attitudes and beliefs. It is harder to formalize and track. It is a diffused depiction of humans and society. It is at the root of all of the other forms of allocative harm.
What can we do to tackle these problems?
Start working on fairness forensics
Test our systems: eg: build pre-release trials to see how a system is working across different populations
How do we track the life cycle of a training dataset to know who built it and what the demographics skews might be in that dataset
Start taking interdisciplinarity seriously
Working with people who are not in our field but have deep expertise in other areas Eg: FATE (Fairness Accountability Transparency Ethics) group at Microsoft Research
Build spaces for collaboration like the AI now institute.
Think harder on the ethics of classification
The ultimate question for fairness in machine learning is this.
Who is going to benefit from the system we are building? And who might be harmed?
Kate Crawford is a Principal Researcher at Microsoft Research and a Distinguished Research Professor at New York University. She has spent the last decade studying the social implications of data systems, machine learning, and artificial intelligence. Her recent publications address data bias and fairness, and social impacts of artificial intelligence among others.
Zhu Long, co-founder and CEO of Yitu Technology, has his identity checked at the company’s headquarters in the Hongqiao business district in Shanghai. Picture: Zigor Aldama
“Our machines can very easily recognise you among at least 2 billion people in a matter of seconds,” says chief executive and Yitu co-founder Zhu Long, “which would have been unbelievable just three years ago.”
Its platform is also in service with more than 20 provincial public security departments, and is used as part of more than 150 municipal public security systems across the country, and Dragonfly Eye has already proved its worth.
On its very first day of operation on the Shanghai Metro, in January, the system identified a wanted man when he entered a station. After matching his face against the database, Dragonfly Eye sent his photo to a policeman, who made an arrest.
In the following three months, 567 suspected lawbreakers were caught on the city’s underground network.
Whole cities in which the algorithms are working say they have seen a decrease in crime. According to Yitu, which says it gets its figures directly from the local authorities, since the system has been implemented, pickpocketing on Xiamen’s city buses has fallen by 30 per cent; 500 criminal cases have been resolved by AI in Suzhou since June 2015; and police arrested nine suspects identified by algorithms during the 2016 G20 summit in Hangzhou.
“Chinese authorities are collecting and centralising ever more information about hundreds of millions of ordinary people, identifying persons who deviate from what they determine to be ‘normal thought’ and then surveilling them,” says Sophie Richardson, China director at HRW.
Research and advocacy group Human Rights Watch (HRW) says security systems such as those being developed by Yitu “violate privacy and target dissent”.
The NGO calls it a “police cloud” system and believes “it is designed to track and predict the activities of activists, dissidents and ethnic minorities, including those authorities say have extreme thoughts, among others”.
Zhu says, “We all discuss AI as an opportunity for humanity to advance or as a threat to it. What I believe is that we will have to redefine what it is to be human. We will have to ask ourselves what the foundations of our species are.
“At the same time, AI will allow us to explore the boundaries of human intelligence, evaluate its performance and help us understand ourselves better.”
Chamath Palihapitiya speaks at a Vanity Fair event in October 2016. Photo by Mike Windle/Getty Images for Vanity Fair
Chamath Palihapitiya, who joined Facebook in 2007 and became its vice president for user growth, said he feels “tremendous guilt” about the company he helped make.
“I think we have created tools that are ripping apart the social fabric of how society works”
Palihapitiya’s criticisms were aimed not only at Facebook, but the wider online ecosystem.
“The short-term, dopamine-driven feedback loops we’ve created are destroying how society works,” he said, referring to online interactions driven by “hearts, likes, thumbs-up.” “No civil discourse, no cooperation; misinformation, mistruth. And it’s not an American problem — this is not about Russians ads. This is a global problem.”
He went on to describe an incident in India where hoax messages about kidnappings shared on WhatsApp led to the lynching of seven innocent people.
“That’s what we’re dealing with,” said Palihapitiya. “And imagine taking that to the extreme, where bad actors can now manipulate large swathes of people to do anything you want. It’s just a really, really bad state of affairs.”
In his talk, Palihapitiya criticized not only Facebook, but Silicon Valley’s entire system of venture capital funding.
He said that investors pump money into “shitty, useless, idiotic companies,” rather than addressing real problems like climate change and disease.
Chamath has not been at Facebook for over six years. When Chamath was at Facebook we were focused on building new social media experiences and growing Facebook around the world. Facebook was a very different company back then and as we have grown we have realised how our responsibilities have grown too. We take our role very seriously and we are working hard to improve. We’ve done a lot of work and research with outside experts and academics to understand the effects of our service on well-being, and we’re using it to inform our product development. We are also making significant investments more in people, technology and processes, and – as Mark Zuckerberg said on the last earnings call – we are willing to reduce our profitability to make sure the right investments are made.
On October 19, the UAE became the first nation with a government minister dedicated to AI. Yes, the UAE now has a minister for artificial intelligence.
“We want the UAE to become the world’s most prepared country for artificial intelligence,” UAE Vice President and Prime Minister and Ruler of Dubai His Highness Sheikh Mohammed bin Rashid Al Maktoum said during the announcement of the position.
The first person to occupy the state minister for AI post is H.E. Omar Bin Sultan Al Olama. The 27-year-old is currently the Managing Director of the World Government Summit in the Prime Minister’s Office at the Ministry of Cabinet Affairs and the Future,
“We have visionary leadership that wants to implement these technologies to serve humanity better. Ultimately, we want to make sure that we leverage that while, at the same time, overcoming the challenges that might be created by AI.” Al Olama
The UAE hopes its AI initiatives will encourage the rest of the world to really consider how our AI-powered future should look.
“AI is not negative or positive. It’s in between. The future is not going to be a black or white. As with every technology on Earth, it really depends on how we use it and how we implement it,”
“People need to be part of the discussion. It’s not one of those things that just a select group of people need to discuss and focus on.”
“At this point, it’s really about starting conversations — beginning conversations about regulations and figuring out what needs to be implemented in order to get to where we want to be. I hope that we can work with other governments and the private sector to help in our discussions and to really increase global participation in this debate.
With regards to AI, one country can’t do everything. It’s a global effort,” Al Olama said.
The prestigious Neural Information Processing Systems conference have a new topic on their agenda. Alongside the usual … concern about AI’s power.
Kate Crawford … urged attendees to start considering, and finding ways to mitigate, accidental or intentional harms caused by their creations. “
“Amongst the very real excitement about what we can do there are also some really concerning problems arising”
“In domains like medicine we can’t have these models just be a black box where something goes in and you get something out but don’t know why,” says Maithra Raghu, a machine-learning researcher at Google. On Monday, she presented open-source software developed with colleagues that can reveal what a machine-learning program is paying attention to in data. Itmay ultimately allow a doctor to see what part of a scan or patient history led an AI assistant to make a particular diagnosis.
“If you have a diversity of perspectives and background you might be more likely to check for bias against different groups” Hanna Wallach a researcher at Microsoft
Others in Long Beach hope to make the people building AI better reflect humanity. Like computer science as a whole, machine learning skews towards the white, male, and western. A parallel technical conference called Women in Machine Learning has run alongside NIPS for a decade. This Friday sees the first Black in AI workshop, intended to create a dedicated space for people of color in the field to present their work.
Towards the end of her talk Tuesday, Crawford suggested civil disobedience could shape the uses of AI. She talked of French engineer Rene Carmille, who sabotaged tabulating machines used by the Nazis to track French Jews. And she told today’s AI engineers to consider the lines they don’t want their technology to cross. “Are there some things we just shouldn’t build?” she asked.
[Timnit] Gebru, 34, joined a Microsoft Corp. team called FATE—for Fairness, Accountability, Transparency and Ethics in AI. The program was set up three years ago to ferret out biases that creep into AI data and can skew results.
“I started to realize that I have to start thinking about things like bias. Even my own Phd work suffers from whatever issues you’d have with dataset bias.”
Companies, government agencies and hospitals are increasingly turning to machine learning, image recognition and other AI tools to help predict everything from the credit worthiness of a loan applicant to the preferred treatment for a person suffering from cancer. The tools have big blind spots that particularly effect women and minorities.
“The worry is if we don’t get this right, we could be making wrong decisions that have critical consequences to someone’s life, health or financial stability,” says Jeannette Wing, director of Columbia University’s Data Sciences Institute.
AI also has a disconcertingly human habit of amplifying stereotypes. Phd students at the University of Virginia and University of Washington examined a public dataset of photos and found that the images of people cooking were 33 percent more likely to picture women than men. When they ran the images through an AI model, the algorithms said women were 68 percent more likely to appear in the cooking photos.
Researchers say it will probably take years to solve the bias problem.
The good news is that some of the smartest people in the world have turned their brainpower on the problem. “The field really has woken up and you are seeing some of the best computer scientists, often in concert with social scientists, writing great papers on it,” says University of Washington computer science professor Dan Weld. “There’s been a real call to arms.”
Training an AI platform on social media data, with the intent to reproduce a “human” experience, is fraught with risk. You could liken it to raising a baby on a steady diet of Fox News or CNN, with no input from its parents or social institutions. In either case, you might be breeding a monster.
Ultimately, social data — alone — represents neither who we actually are nor who we should be. Deeper still, as useful as the social graph can be in providing a training set for AI, what’s missing is a sense of ethics or a moral framework to evaluate all this data. From the spectrum of human experience shared on Twitter, Facebook and other networks, which behaviors should be modeled and which should be avoided? Which actions are right and which are wrong? What’s good … and what’s evil?
Here’s where science comes up short. The answers can’t be gleaned from any social data set. The best analytical tools won’t surface them, no matter how large the sample size.
But they just might be found in the Bible. And the Koran, the Torah, the Bhagavad Gita and the Buddhist Sutras. They’re in the work of Aristotle, Plato, Confucius, Descartes and other philosophers both ancient and modern.
AI, to be effective, needs an ethical underpinning. Data alone isn’t enough. AI needs religion — a code that doesn’t change based on context or training set.
In place of parents and priests, responsibility for this ethical education will increasingly rest on frontline developers and scientists.
As emphasized by leading AI researcher Will Bridewell, it’s critical that future developers are “aware of the ethical status of their work and understand the social implications of what they develop.” He goes so far as to advocate study in Aristotle’s ethics and Buddhist ethics so they can “better track intuitions about moral and ethical behavior.”
On a deeper level, responsibility rests with the organizations that employ these developers, the industries they’re part of, the governments that regulate those industries and — in the end — us.
Source: Recode Ryan Holmes is the founder and CEO of Hootsuite.
Donald Trump meeting PayPal co-founder Peter Thiel and Apple CEO Tim Cook in December last year. Photograph: Evan Vucci/AP
One of the biggest puzzles about our current predicament with fake news and the weaponisation of social media is why the folks who built this technology are so taken aback by what has happened.
We have a burgeoning genre of “OMG, what have we done?” angst coming from former Facebook and Google employees who have begun to realise that the cool stuff they worked on might have had, well, antisocial consequences.
Put simply, what Google and Facebook have built is a pair of amazingly sophisticated, computer-driven engines for extracting users’ personal information and data trails, refining them for sale to advertisers in high-speed data-trading auctions that are entirely unregulated and opaque to everyone except the companies themselves.
The purpose of this infrastructure was to enable companies to target people with carefully customised commercial messages and, as far as we know, they are pretty good at that.
It never seems to have occurred to them that their advertising engines could also be used to deliver precisely targeted ideological and political messages to voters. Hence the obvious question: how could such smart people be so stupid?
My hunch is it has something to do with their educational backgrounds. Take the Google co-founders. Sergey Brin studied mathematics and computer science. His partner, Larry Page, studied engineering and computer science.Zuckerberg dropped out of Harvard, where he was studying psychology and computer science, but seems to have been more interested in the latter.
Now mathematics, engineering and computer science are wonderful disciplines – intellectually demanding and fulfilling. And they are economically vital for any advanced society. But mastering them teaches students very little about society or history – or indeed about human nature.
As a consequence, the new masters of our universe are people who are essentially only half-educated. They have had no exposure to the humanities or the social sciences, the academic disciplines that aim to provide some understanding of how society works, of history and of the roles that beliefs, philosophies, laws, norms, religion and customs play in the evolution of human culture.
We are now beginning to see the consequences of the dominance of this half-educated elite.
Source: The Gaurdian – John Naughton is professor of the public understanding of technology at the Open University.
A discussion between OpenAI Director Shivon Zilis and AI Fund Director of Ethics and Governance Tim Hwang, and both shared perspective on AI’s progress, its public perception, and how we can help ensure its responsible development going forward.
Hwang brought up the fact that artificial intelligence researchers are, in some ways, “basically writing policy in code”because of how influential the particular perspectives or biases inherent in these systems will be, and suggested that researchers could actually consciously set new cultural norms via their work.
Zilis added that the total number of people setting the tone for incredibly intelligent AI is probably “in the low thousands.”
She added that this means we likely need more crossover discussion between this community and those making policy decisions, and Hwang added that currently, there’s
“no good way for the public at large to signal” what moral choices should be made around the direction of AI development.
Zilis concluded that she has three guiding principles in terms of how she thinks about the future of responsible artificial intelligence development:
First, the tech’s coming no matter what, so we need to figure out how to bend its arc with intent.
Second, how do we get more people involved in the conversation?
And finally, we need to do our best to front load the regulation and public discussion needed on the issue, since ultimately, it’s going to be a very powerful technology.
From left: Twitter’s acting general counsel Sean Edgett, Facebook’s general counsel Colin Stretch and Google’s senior vice president and general counsel Kent Walker, testify before the House Intelligence Committee on Wednesday, Nov. 1, 2017. Manuel Balce Ceneta/AP
Members of Congress confessed how difficult it was for them to even wrap their minds around how today’s Internet works — and can be abused. And for others, the hearings finally drove home the magnitude of the Big Tech platforms.
Sen. John Kennedy, R-La., marveled on Tuesday when Facebook said it could track the source of funding for all 5 million of its monthly advertisers.
“I think you do enormous good, but your power scares me,” he said.
There appears to be no quick patch for the malware afflicting America’s political life.
Over the course of three congressional hearings Tuesday and Wednesday, lawmakers fulminated, Big Tech witnesses were chastened but no decisive action appears to be in store to stop a foreign power from harnessing digital platforms to try to shape the information environment inside the United States.
Legislation offered in the Senate — assuming it passed, months or more from now — would change the calculus slightly: requiring more disclosure and transparency for political ads on Facebook and Twitter and other social platforms.
Even if it became law, however, it would not stop such ads from being sold, nor heal the deep political divisions exploited last year by foreign influence-mongers. The legislation also couldn’t stop a foreign power from using all the other weapons in its arsenal against the U.S., including cyberattacks, the deployment of human spies and others.
“Candidly, your companies know more about Americans, in many ways, than the United States government does. The idea that you had no idea any of this was happening strains my credibility,”Senate Intelligence Committee Vice Chairman Mark Warner, D.-Va.
The companies also made clear they condemn the uses of their services they’ve discovered, which they said violate their policies in many cases.
They also talked more about the scale of the Russian digital operation they’ve uncovered up to this point — which is eye-watering: Facebook general counsel Colin Stretch acknowledged that as many as 150 million Americans may have seen posts or other content linked to Russia’s influence campaign in the 2016 cycle
“There is one thing I’m certain of, and it’s this: Given the complexity of what we have seen, if anyone tells you they have figured it out, they are kidding ourselves. And we can’t afford to kid ourselves about what happened last year — and continues to happen today.” Senate Intelligence Committee Chairman Richard Burr, R-N.C.
In a statement broadcast live on Facebook on September 21 and subsequently posted to his profile page, Zuckerberg pledged to increase the resources of Facebook’s security and election-integrity teams and to work “proactively to strengthen the democratic process.”
It was an admirable commitment. But reading through it, I kept getting stuck on one line: “We have been working to ensure the integrity of the German elections this weekend,” Zuckerberg writes. It’s a comforting sentence, a statement that shows Zuckerberg and Facebook are eager to restore trust in their system.
But … it’s not the kind of language we expect from media organizations, even the largest ones. It’s the language of governments, or political parties, or NGOs. A private company, working unilaterally to ensure election integrity in a country it’s not even based in?
Facebook has grown so big, and become so totalizing, that we can’t really grasp it all at once.
Like a four-dimensional object, we catch slices of it when it passes through the three-dimensional world we recognize. In one context, it looks and acts like a television broadcaster, but in this other context, an NGO. In a recent essay for the London Review of Books, John Lanchester argued that for all its rhetoric about connecting the world, the company is ultimately built to extract data from users to sell to advertisers. This may be true, but Facebook’s business model tells us only so much about how the network shapes the world.
Between March 23, 2015, when Ted Cruz announced his candidacy, and November 2016, 128 million people in America created nearly 10 billion Facebook posts, shares, likes, and comments about the election. (For scale, 137 million people voted last year.)
In February 2016, the media theorist Clay Shirky wrote about Facebook’s effect: “Reaching and persuading even a fraction of the electorate used to be so daunting that only two national orgs” — the two major national political parties — “could do it. Now dozens can.”
It used to be if you wanted to reach hundreds of millions of voters on the right, you needed to go through the GOP Establishment. But in 2016, the number of registered Republicans was a fraction of the number of daily American Facebook users, and the cost of reaching them directly was negligible.
Tim Wu, the Columbia Law School professor
“Facebook has the same kind of attentional power [as TV networks in the 1950s], but there is not a sense of responsibility,” he said. “No constraints. No regulation. No oversight. Nothing. A bunch of algorithms, basically, designed to give people what they want to hear.”
It tends to get forgotten, but Facebook briefly ran itself in part as a democracy: Between 2009 and 2012, users were given the opportunity to vote on changes to the site’s policy. But voter participation was minuscule, and Facebook felt the scheme “incentivized the quantity of comments over their quality.” In December 2012, that mechanism was abandoned “in favor of a system that leads to more meaningful feedback and engagement.”
Facebook had grown too big, and its users too complacent, for democracy.
As the director of Stanford’s AI Lab and now as a chief scientist of Google Cloud, Fei-Fei Li is helping to spur the AI revolution. But it’s a revolution that needs to include more people. She spoke with MIT Technology Review senior editor Will Knight about why everyone benefits if we emphasize the human side of the technology.
Why did you join Google?
Researching cutting-edge AI is very satisfying and rewarding, but we’re seeing this great awakening, a great moment in history. For me it’s very important to think about AI’s impact in the world, and one of the most important missions is to democratize this technology. The cloud is this gigantic computing vehicle that delivers computing services to every single industry.
What have you learned so far?
We need to be much more human-centered.
If you look at where we are in AI, I would say it’s the great triumph of pattern recognition. It is very task-focused, it lacks contextual awareness, and it lacks the kind of flexible learning that humans have.
We also want to make technology that makes humans’ lives better, our world safer, our lives more productive and better. All this requires a layer of human-level communication and collaboration.
When you are making a technology this pervasive and this important for humanity, you want it to carry the values of the entire humanity, and serve the needs of the entire humanity.
If the developers of this technology do not represent all walks of life, it is very likely that this will be a biased technology. I say this as a technologist, a researcher, and a mother. And we need to be speaking about this clearly and loudly.
The unit, called DeepMind Ethics and Society, is not the AI Ethics Board that DeepMind was promised when it agreed to be acquired by Google in 2014. That board, which was convened by January 2016, was supposed to oversee all of the company’s AI research, but nothing has been heard of it in the three-and-a-half years since the acquisition. It remains a mystery who is on it, what they discuss, or even whether it has officially met.
DeepMind Ethics and Society is also not the same as DeepMind Health’s Independent Review Panel, a third body set up by the company to provide ethical oversight – in this case, of its specific operations in healthcare.
Nor is the new research unit the Partnership on Artificial Intelligence to Benefit People and Society, an external group founded in part by DeepMind and chaired by the company’s co-founder Mustafa Suleyman. That partnership, which was also co-founded by Facebook, Amazon, IBM and Microsoft, exists to “conduct research, recommend best practices, and publish research under an open licence in areas such as ethics, fairness and inclusivity”.
Nonetheless, its creation is the hallmark of a change in attitude from DeepMind over the past year,which has seen the company reassess its previously closed and secretive outlook. It is still battling a wave of bad publicity started when it partnered with the Royal Free in secret, bringing the app Streams to active use in the London hospital without being open to the public about what data was being shared and how.
The research unit also reflects an urgency on the part of many AI practitioners to get ahead of growing concerns on the part of the public about how the new technology will shape the world around us.
We believe AI can be of extraordinary benefit to the world, but only if held to the highest ethical standards.
Technology is not value neutral, and technologists must take responsibility for the ethical and social impact of their work.
As history attests, technological innovation in itself is no guarantee of broader social progress. The development of AI creates important and complex questions. Its impact on society—and on all our lives—is not something that should be left to chance. Beneficial outcomes and protections against harms must be actively fought for and built-in from the beginning. But in a field as complex as AI, this is easier said than done.
As scientists developing AI technologies, we have a responsibility to conduct and support open research and investigation into the wider implications of our work. At DeepMind, we start from the premise that all AI applications should remain under meaningful human control, and be used for socially beneficial purposes.
So today we’re launching a new research unit, DeepMind Ethics & Society, to complement our work in AI science and application. This new unit will help us explore and understand the real-world impacts of AI. It has a dual aim: to help technologists put ethics into practice, and to help society anticipate and direct the impact of AI so that it works for the benefit of all.
If AI technologies are to serve society, they must be shaped by society’s priorities and concerns.
The missteps underscore how misinformation continues to undermine the credibility of Silicon Valley’s biggest companies.
Accuracy matters in the moments after a tragedy. Facts can help catch the suspects, save lives and prevent a panic.
But in the aftermath of the deadly mass shooting in Las Vegas on Sunday, the world’s two biggest gateways for information, Google and Facebook, did nothing to quell criticism that they amplify fake news when they steer readers toward hoaxes and misinformation gathering momentum on fringe sites.
Google posted under its “top stories” conspiracy-laden links from 4chan — home to some of the internet’s most ardent trolls. It also promoted a now-deleted story from Gateway Pundit and served videos on YouTube of dubious origin.
The posts all had something in common: They identified the wrong assailant.
Facebook’s Crisis Response page, a hub for users to stay informed and mobilize during disasters, perpetuated the same rumors by linking to sites such as Alt-Right News and End Time Headlines, according to Fast Company.
The platforms have immense influence on what gets seen and read. More than two-thirds of Americans report getting at least some of their news from social media, according to the Pew Research Center. A separate global study published by Edelman last year found that more people trusted search engines (63%) for news and information than traditional media such as newspapers and television (58%).
Still, skepticism abounds that the companies beholden to shareholders are equipped to protect the public from misinformation and recognize the threat their platforms pose to democratic societies.
I was asking in the context of the aftermath of the 2016 election and the misinformation that companies like Facebook, Twitter, and Google were found to have spread.
“I view it as a big responsibility to get it right,” he says. “I think we’ll be able to do these things better over time. But I think the answer to your question, the short answer and the only answer, is we feel huge responsibility.”
But it’s worth questioning whether Google’s systems are making the rightdecisions, even as they make some decisions much easier.
People are already skittish about how much Google knows about them, and they are unclear on how to manage their privacy settings. Pichai thinks that’s another one of those problems that AI could fix, “heuristically.”
“Down the line, the system can be much more sophisticated about understanding what is sensitive for users, because it understands context better,” Pichai says. “[It should be] treating health-related information very differently from looking for restaurants to eat with friends.” Instead of asking users to sift through a “giant list of checkboxes,” a user interface driven by AI could make it easier to manage.
Of course, what’s good for users versus what’s good for Google versus what’s good for the other business that rely on Google’s data is a tricky question. And it’s one that AI alone can’t solve. Google is responsible for those choices, whether they’re made by people or robots.
The amount of scrutiny companies like Facebook and Google — and Google’s YouTube division — face over presenting inaccurate or outright manipulative information is growing every day, and for good reason.
Pichai thinks that Google’s basic approach for search can also be used for surfacing good, trustworthy content in the feed. “We can still use the same core principles we use in ranking around authoritativeness, trust, reputation.
What he’s less sure about, however, is what to do beyond the realm of factual information — with genuine opinion: “I think the issue we all grapple with is how do you deal with the areas where people don’t agree or the subject areas get tougher?”
When it comes to presenting opinions on its feed, Pichai wonders if Google could “bring a better perspective, rather than just ranking alone. … Those are early areas of exploration for us, but I think we could do better there.”
John Giannandrea, who leads AI at Google, is worried about intelligent systems learning human prejudices.
… concerned about the danger that may be lurking inside the machine-learning algorithms used to make millions of decisions every minute.
The real safety question, if you want to call it that, is that if we give these systems biased data, they will be biased
The problem of bias in machine learning is likely to become more significant as the technology spreads to critical areas like medicine and law, and as more people without a deep technical understanding are tasked with deploying it. Some experts warn that algorithmic bias is already pervasive in many industries, and that almost no one is making an effort to identify or correct it.
Karrie Karahalios, a professor of computer science at the University of Illinois, presented research highlighting how tricky it can be to spot bias in even the most commonplace algorithms. Karahalios showed that users don’t generally understand how Facebook filters the posts shown in their news feed. While this might seem innocuous, it is a neat illustration of how difficult it is to interrogate an algorithm.
Facebook’s news feed algorithm can certainly shape the public perception of social interactions and even major news events. Other algorithms may already be subtly distorting the kinds of medical care a person receives, or how they get treated in the criminal justice system.
This is surely a lot more important than killer robots, at least for now.
“Personally, I think the idea that fake news on Facebook, it’s a very small amount of the content, influenced the election in any way is a pretty crazy idea.”
Facebook CEO Mark Zuckerberg’s company recently said it would turn over to Congress more than 3,000 politically themed advertisements that were bought by suspected Russian operatives. (Eric Risberg/AP
Nine days after Facebook chief executive Mark Zuckerberg dismissed as “crazy” the idea that fake news on his company’s social network played a key role in the U.S. election, President Barack Obama pulled the youthful tech billionaire aside and delivered what he hoped would be a wake-up call.
Obama made a personal appeal to Zuckerberg to take the threat of fake news and political disinformation seriously. Unless Facebook and the government did more to address the threat, Obama warned, it would only get worse in the next presidential race.
“There’s been a systematic failure of responsibility. It’s rooted in their overconfidence that they know best, their naivete about how the world works, their extensive effort to avoid oversight, and their business model of having very few employees so that no one is minding the store.”Zeynep Tufekci
Zuckerberg acknowledged the problem posed by fake news. But he told Obama that those messages weren’t widespread on Facebook and that there was no easy remedy, according to people briefed on the exchange
One outcome of those efforts was Zuckerberg’s admission on Thursday that Facebook had indeed been manipulated and that the company would now turn over to Congress more than 3,000 politically themed advertisements that were bought by suspected Russian operatives.
These issues have forced Facebook and other Silicon Valley companies to weigh core values, including freedom of speech, against the problems created when malevolent actors use those same freedoms to pump messages of violence, hate and disinformation.
Congressional investigators say the disclosure only scratches the surface. One called Facebook’s discoveries thus far “the tip of the iceberg.” Nobody really knows how many accounts are out there and how to prevent more of them from being created to shape the next election — and turn American society against itself.
“There is no question that the idea that Silicon Valley is the darling of our markets and of our society — that sentiment is definitely turning,” said Tim O’Reilly, an adviser to tech executives and chief executive of the influential Silicon Valley-based publisher O’Reilly Media
IBM Chairman, President, and Chief Executive Officer Ginni Rometty. PHOTOGRAPHER: STEPHANIE SINCLAIR FOR BLOOMBERG BUSINESSWEEK
If I considered the initials AI, I would have preferred augmented intelligence.
It’s the idea that each of us are going to need help on all important decisions.
A study said on average that a third of your decisions are really great decisions, a third are not optimal, and a third are just wrong. We’ve estimated the market is $2 billion for tools to make better decisions.
That’s what led us all to really calling it cognitive
“Look, we really think this is about man and machine, not man vs. machine. This is an era—really, an era that will play out for decades in front of us.”
We set out to build an AI platform for business.
AI would be vertical. You would train it to know medicine. You would train it to know underwriting of insurance. You would train it to know financial crimes. Train it to know oncology. Train it to know weather. And it isn’t just about billions of data points. In the regulatory world, there aren’t billions of data points. You need to train and interpret something with small amounts of data.
This is really another key point about professional AI. Doctors don’t want black-and-white answers, nor does any profession. If you’re a professional, my guess is when you interact with AI, you don’t want it to say, “Here is an answer.”
What a doctor wants is, “OK, give me the possible answers. Tell my why you believe it. Can I see the research, the evidence, the ‘percent confident’? What more would you like to know?”
It’s our responsibility if we build this stuff to guide it safely into the world.
“People have serious conversations with Siri. People talk to Siri about all kinds of things, including when they’re having a stressful day or have something serious on their mind. They turn to Siri in emergencies or when they want guidance on living a healthier life. Does improving Siri in these areas pique your interest?
Come work as part of the Siri Domains team and make a difference.
We are looking for people passionate about the power of data and have the skills to transform data to intelligent sources that will take Siri to next level. Someone with a combination of strong programming skills and a true team player who can collaborate with engineers in several technical areas. You will thrive in a fast-paced environment with rapidly changing priorities.”
The challenge as explained by Ephrat Livni on Quartz
The position requires a unique skill set. Basically, the company is looking for a computer scientist who knows algorithms and can write complex code, but also understands human interaction, has compassion, and communicates ably, preferably in more than one language. The role also promises a singular thrill: to “play a part in the next revolution in human-computer interaction.”
The job at Apple has been up since April, so maybe it’s turned out to be a tall order to fill. Still, it shouldn’t be impossible to find people who are interested in making machines more understanding. If it is, we should probably stop asking Siri such serious questions.
Computer scientists developing artificial intelligence have long debated what it means to be human and how to make machines more compassionate. Apart from the technical difficulties, the endeavor raises ethical dilemmas, as noted in the 2012 MIT Press book Robot Ethics: The Ethical and Social Implications of Robotics.
Even if machines could be made to feel for people, it’s not clear what feelings are the right ones to make a great and kind advisor and in what combinations. A sad machine is no good, perhaps, but a real happy machine is problematic, too.
In a chapter on creating compassionate artificial intelligence (pdf), sociologist, bioethicist, and Buddhist monk James Hughes writes:
Programming too high a level of positive emotion in an artificial mind, locking it into a heavenly state of self-gratification, would also deny it the capacity for empathy with other beings’ suffering, and the nagging awareness that there is a better state of mind.
Geoffrey Hinton harbors doubts about AI’s current workhorse. (Johnny Guatto / University of Toronto)
In 1986, Geoffrey Hinton co-authored a paper that, three decades later, is central to the explosion of artificial intelligence.
But Hinton says his breakthrough method should be dispensed with, and a new path to AI found.
… he is now “deeply suspicious” of back-propagation, the workhorse method that underlies most of the advances we are seeing in the AI field today, including the capacity to sort through photos and talk to Siri.
“My view is throw it all away and start again”
Hinton said that, to push materially ahead, entirely new methods will probably have to be invented. “Max Planck said, ‘Science progresses one funeral at a time.’ The future depends on some graduate student who is deeply suspicious of everything I have said.”
Hinton suggested that, to get to where neural networks are able to become intelligent on their own, what is known as “unsupervised learning,” “I suspect that means getting rid of back-propagation.”
“I don’t think it’s how the brain works,” he said. “We clearly don’t need all the labeled data.”
PL – The challenge of an AI using Emotion Reading Tech just got dramatically more difficult.
A new study identifies 27 categories of emotion and shows how they blend together in our everyday experience.
Psychology once assumed that most human emotions fall within the universal categories of happiness, sadness, anger, surprise, fear, and disgust. But a new study from Greater Good Science Center faculty director Dacher Keltner suggests that there are at least 27 distinct emotions—and they are intimately connected with each other.
“We found that 27 distinct dimensions, not six, were necessary to account for the way hundreds of people reliably reported feeling in response to each video”
Moreover, in contrast to the notion that each emotional state is an island, the study found that “there are smooth gradients of emotion between, say, awe and peacefulness, horror and sadness, and amusement and adoration,”Keltner said.
“We don’t get finite clusters of emotions in the map because everything is interconnected,” said study lead author Alan Cowen, a doctoral student in neuroscience at UC Berkeley.
“Emotional experiences are so much richer and more nuanced than previously thought.”
…or the learned morals of an evolving algorithm. SAS CTO Oliver Schabenberger With the advent of deep learning, machines are beginning to solve problems in a novel way: by writing the algorithms themselves.
The software developer who codifies a solution through programming logic is replaced by a data scientist who defines and trains a deep neural network.
The expert who studied and learned a domain is replaced by a reinforcement learning algorithm that discovers the rules of play from historical data.
We are learning incredible lessons in this process.
But does the rise of such highly sophisticated deep learning mean that machines will soon surpass their makers? They are surpassing us in reliability, accuracy and throughput. But they are not surpassing us in thinking or learning.Not with today’s technology.
The artificial intelligence systems of today learn from data – they learn only from data. These systems cannot grow beyond the limits of the data by creating, innovating or reasoning.
Even a reinforcement learning system that discovers rules of play from past data cannot develop completely new rules or new games. It can apply the rules in a novel and more efficient way, but it does not invent a new game. The machine that learned to play Go better than any human being does not know how to play Poker.
Where to from here?
True intelligence requires creativity, innovation, intuition, independent problem solving, self-awareness and sentience. The systems built based on deep learning do not – and cannot – have these characteristics.These are trained by top-down supervised methods.
We first tell the machine the ground truth, so that it can discover its regularities. They do not grow beyond that.
AI works, in part, because complex algorithms adeptly identify, remember, and relate data … Moreover, some machines can do what had been the exclusive domain of humans and other intelligent life:Learn on their own.
As a researcher schooled inscientific methodandan ethicist immersed in moral decision-making, I know it’s challenging for humans to navigate concurrently the two disparate arenas.
It’s even harder to envision how computer algorithms can enable machines to act morally.
Moral choice, however, doesn’t ask whether an action will produce an effective outcome; it asks if it isa good decision. In other words, regardless of efficacy, is it the right thing to do?
Such analysis does not reflect an objective, data-driven decision but a subjective, judgment-based one.
Individuals often make moral decisions on the basis of principles like decency, fairness, honesty, and respect. To some extent, people learn those principles through formal study and reflection; however, the primary teacher is life experience, which includes personal practice and observation of others.
Placing manipulative ads before a marginally-qualified and emotionally vulnerable target market may be very effective for the mortgage company, but many people would challenge the promotion’s ethicality.
Humans can make that moral judgment, but how does a data-driven computer draw the same conclusion? Therein lies what should be a chief concern about AI.
Can computers be manufactured with a sense of decency?
Can coding incorporate fairness? Can algorithms learn respect?
It seems incredible for machines to emulate subjective, moral judgment, but if that potential exists, at least four critical issues must be resolved:
SAN JOSE, CA – APRIL 18: Facebook CEO Mark Zuckerberg delivers the keynote address at Facebook’s F8 Developer Conference on April 18, 2017 at McEnery Convention Center in San Jose, California. (Photo by Justin Sullivan/Getty Images)
… recent story in The Washington Post reported that “minority” groups feel unfairly censored by social media behemoth Facebook, for example, when using the platform for discussions about racial bias. At the same time, groups and individuals on the other end of the race spectrum are quickly being banned and ousted in a flash from various social media networks.
Most all of such activity begins with an algorithm, a set of computer code that, for all intents and purposes for this piece, is created to raise a red flag when certain speech is used on a site.
But from engineer mindset to tech limitation, just how much faith should we be placing in algorithms when it comes to the very sensitive area of digital speech and race, and what does the future hold?
Indeed, while Facebook head Mark Zuckerberg reportedly eyes political ambitions within an increasingly brown America in which his own company consistently has issues creating racial balance, there are questions around policy and development of such algorithms. In fact, Malkia Cyril executive director for the Center for Media Justice told the Post that she believes that Facebook has a double standard when it comes to deleting posts.
Cyril explains [her meeting with Facebook] “The meeting was a good first step, but very little was done in the direct aftermath. Even then, Facebook executives, largely white, spent a lot of time explaining why they could not do more instead of working with us to improve the user experience for everyone.”
What’s actually in the hearts and minds of those in charge of the software development? How many more who are coding have various thoughts – or more extreme – as those recently expressed in what is now known as the Google Anti-Diversity memo?
Not just Facebook, but any and all tech platforms where race discussion occurs are seemingly at a crossroads and under various scrutiny in terms of management, standards and policy about this sensitive area. The main question is how much of this imbalance is deliberate and how much is just a result of how algorithms naturally work?
Nelson [National Chairperson National Society of Black Engineers] notes that the first source of error, however, is how a particular team defines the term hate speech. “That opinion may differ between people so any algorithm would include error at the individual level,” he concludes.
“I believe there are good people at Facebook who want to see justice done,” says Cyril. “There are steps being taken at the company to improve the experience of users and address the rising tide of hate that thwarts democracy, on social media and in real life.
That said, racism is not race neutral, and accountability for racism will never come from an algorithm alone.”
Two prominent research-image collections—including one supported by Microsoft and Facebook—display a predictable gender bias in their depiction of activities such as cooking and sports. Images of shopping and washing are linked to women, for example, while coaching and shooting are tied to men.
Machine-learning software trained on the datasets didn’t just mirror those biases, it amplified them. If a photo set generally associated women with cooking, software trained by studying those photos and their labels created an even stronger association.
Mark Yatskar, a researcher at the Allen Institute for Artificial Intelligence, says that this phenomenon could also amplify other biases in data, for example related to race. “This could work to not only reinforce existing social biases but actually make them worse,”says Yatskar, who worked with Ordóñez and others on the project while at the University of Washington.
“A system that takes action that can be clearly attributed to gender bias cannot effectively function with people,” he says.
When image-recognition software is “trained” by examining these datasets, the bias is amplified. A system trained on the COCO dataset associated men with keyboards and computer mice even more strongly than the dataset itself.
The researchers devised a way to neutralize this amplification phenomenon—effectively forcing learning software to reflect its training data. But it requires a researcher to be looking for bias in the first place, and to specify what he or she wants to correct. And the corrected software still reflects the gender biases baked into the original data.
One point of agreement in the field is that using machine learning to solve problems is more complicated than many people previously thought.
“Work like this is correcting the illusion that algorithms can be blindly applied to solve problems,” says Suresh Venkatasubramanian, a professor at the University of Utah.
Fear of opaque power of Google in particular, and Silicon Valley in general, wields over our lives.
If Google — and the tech world more generally — is sexist, or in the grips of a totalitarian cult of political correctness, or a secret hotbed of alt-right reactionaries, the consequences would be profound.
Google wields a monopoly over search, one of the central technologies of our age, and, alongside Facebook, dominates the internet advertising market, making it a powerful driver of both consumer opinion and the media landscape.
It shapes the world in which we live in ways both obvious and opaque.
This is why trust matters so much in tech. It’s why Google, to attain its current status in society, had to promise, again and again, that it wouldn’t be evil.
Compounding the problem is that the tech industry’s point of view is embedded deep in the product, not announced on the packaging. Its biases are quietly built into algorithms, reflected in platform rules, expressed in code few of us can understand and fewer of us will ever read.
But what if it actually is evil? Or whatif it’s not evil but just immature, unreflective, and uncompassionate?And what if that’s the culture that designs the digital services the rest of us have to use?
The technology industry’s power is vast, and the way that power is expressed is opaque, so the only real assurance you can have that your interests and needs are being considered is to be in the room when the decisions are made and the code is written. But tech as an industry is unrepresentative of the people it serves and unaccountable in the way it serves them, and so there’s very little confidence among any group that the people in the room are the right ones.
If you look at almost every other tool that has ever been created, our tools tend to be most valuable when they’re amplifying us, when they’re extending our reach, when they’re increasing our strength, when they’re allowing us to do things that we can’t do by ourselves as human beings.That’s really the way that we need to be thinking about AI as well,and to the extent that we actually call it augmented intelligence, not artificial intelligence.
Some time ago we realized that this thing called cognitive computing was really bigger than us, it was bigger than IBM, it was bigger than any one vendor in the industry, it was bigger than any of the one or two different solution areas that we were going to be focused on, and we had to open it up, which is when we shifted from focusing on solutions to really dealing with more of a platform of services, where each service really is individually focused on a different part of the problem space.
… what we’re talking about now are a set of services, each of which do something very specific, each of which are trying to deal with a different part of our human experience, and with the idea that anybody building an application, anybody that wants to solve a social or consumer or business problem can do that by taking our services, then composing that into an application.
… If the doctor can now make decisions that are more informed, that are based on real evidence, that are supported by the latest facts in science, that are more tailored and specific to the individual patient, it allows them to actually do their job better. For radiologists, it may allow them to see things in the image that they might otherwise miss or get overwhelmed by. It’s not about replacing them. It’s about helping them do their job better.
That’s really the way to think about this stuff, is that it will have its greatest utility when it is allowing us to do what we do better than we could by ourselves, when the combination of the human and the tool together are greater than either one of them would’ve been by theirselves. That’s really the way we think about it. That’s how we’re evolving the technology. That’s where the economic utility is going to be.
There are lots of things that we as human beings are good at. There’s also a lot of things that we’re not very good, and that’s I think where cognitive computing really starts to make a huge difference, is when it’s able to bridge that distance to make up that gap
A way I like to say it is it doesn’t do our thinking for us, it does our research for us so we can do our thinking better,and that’s true of us as end users and it’s true of advisors.
Those words got to me 18 years ago during an interview I had with this esteemed artist. We were working on a project together, an interactive CD about his movie posters, several of which were classics by then, when the conversation wandered off the subject of art and we began to examine the importance of being true to one’s self.
“Have you ever, in your classes or seminars talked much about the underlying core foundation principles of your life?” I asked Drew that day.
His answer in part went like this: “Whenever I talk, I’m asked to talk about my art, because that’s what theysee, that’s what’s out front. But the power of the art comes out of the personality of the human being. Inevitably, you can’t paint what you ain’t.”
That conversation between us took place five days before Columbine, in April of 1999, when Pam and I lived in Denver and a friend of ours had children attending that school. That horrific event triggered a lot of value discussions and a lot of human actions, in response to it.
Flash-forward to Charlottesville.And an email, in response to it, that the CEO of a large tech companysent his employees yesterday, putting a stake in the ground about what his company stands for, and won’t stand for,during these “horrific” times.
“… At Microsoft, we strive to seek out differences, celebrate them and invite them in. As a leader, a key part of your role is creating a culture where every person can do their best work, which requires more than tolerance for diverse perspectives. Our growth mindset culture requires us to truly understand and share the feelings of another person. …”
If Satya Nadella’s email expresses the emerging personality at Microsoft, the power source from which it works, then we are cautiously optimistic about what this could do for socializing AI.
It will take this kind of foundation-building, going forward, as MS introduces more AI innovations, to diminish the inherent bias in deep learning approaches and the implicit bias in algorithms.
It will take this depth of awareness to shape the values of Human-AI collaboration, to protect the humans who use AI. Values that, “seek out differences, celebrate them and invite them in.”
It will require unwavering dedication to this goal. Because. You can’t paint what you ain’t.
Yesterday (Aug. 14), Microsoft CEO Satya Nadella sent out the following email to employees at Microsoft after the deadly car crash at a white nationalist rally in in Charlottesville, Virginia, on Saturday, Aug. 12:
This past week and in particular this weekend’s events in Charlottesville have been horrific. What I’ve seen and read has had a profound impact on me and I am sure for many of you as well. In these times, to me only two things really matter as a leader.
The first is that we stand for our timeless values, which include diversity and inclusion. There is no place in our society for the bias, bigotry and senseless violence we witnessed this weekend in Virginia provoked by white nationalists. Our hearts go out to the families and everyone impacted by the Charlottesville tragedy.
The second is that we empathize with the hurt happening around us. At Microsoft, we strive to seek out differences, celebrate them and invite them in. As a leader, a key part of your role is creating a culture where every person can do their best work, which requires more than tolerance for diverse perspectives. Our growth mindset culture requires us to truly understand and share the feelings of another person. It is an especially important time to continue to be connected with people, and listen and learn from each other’s experiences.
As I’ve said, across Microsoft, we will stand together with those who are standing for positive change in the communities where we live, work and serve. Together, we must embrace our shared humanity, and aspire to create a society that is filled with respect, empathy and opportunity for all.
Technology and the law are converging, and where they meet new questions arise about the relative roles of artificial and human agents—and the ethical issues involved in the shift from one to the other. While legal technology has largely focused on the activities of the bar, it challenges us to think about its application to the bench as well. In particular,
Could AI replace human judges?
The idea of AI judges raises important ethical issues around bias and autonomy. AI programs may incorporate the biases of their programmers and the humans they interact with.
But while such programs may replicate existing human biases, the distinguishing feature of AI over an algorithm is that it can behave in surprising and unintended ways as it ‘learns.’ Eradicating bias therefore becomes even more difficult, though not impossible. Any AI judging program would need to account for, and be tested for, these biases.
Appealing to rationality, the counter-argument is that human judges are already biased, and that AI can be used to improve the way we deal with them and reduce our ignorance. Yet suspicions about AI judges remain, and are already enough of a concern to lead the European Union to promulgate a General Data Protection Regulation which becomes effective in 2018. This Regulation contains
“the right not to be subject to a decision based solely on automated processing”.
As the English utilitarian legal theorist Jeremy Bentham once wrote in An Introduction To The Principles of Morals and Legislation, “in principle and in practice, in a right track and in a wrong one, the rarest of all human qualities is consistency.” With the ability to process far more data and variables in the case record than humans could ever do, an AI judge might be able to outstrip a human one in many cases.
Even so, AI judges may not solve classical questions of legal validity so much as raise new questions about the role of humans, since—if we believe that ethics and morality in the law are important—then they necessarily lie, or ought to lie, in the domain of human judgment.
In practical terms, if we apply this conclusion to the perspective of American legal theorist Ronald Dworkin, for example, AI could assist with examining the entire breadth and depth of the law, but humans would ultimately choose what they consider a morally-superior interpretation.
The American Judge Richard Posner believes that the immediate use of AI and automation should be restricted to assisting judges in uncovering their own biases and maintaining consistency.
At the heart of these issues is a hugely challenging question: what does it mean to be human in the age of Artificial Intelligence?
Random International’s Zoological, part of Wayne McGregor’s +/- Human Photograph: Ravi Deepres/Alicia Clarke
Random International’s installation, Zoological, features a flock of airborne spheres that glide and swoop and dance and swarm above and among us. What a mind-boggling show.
In the darkened heights of the Roundhouse in north London, a flying flock of white spheres that uncannily resemble Magritte’s dream objects float intelligently and curiously, checking out the humans below, hovering downward to see us better. They are the most convincing embodiment of artificial intelligence I have ever seen. For these responsive, even sensitive machines truly create a sense of encounter with a digital life form that mirrors, or mocks, human free will.
Nobody is hidden behind a screen piloting this robotic airborne dance troupe. Each sphere has its own decision-making electronic brain. They fly in elegant unison yet also break ranks as they check their positions against the images recorded by infra-red cameras surrounding the circular space where they float and their human visitors walk.
Yet the crucial fact that they guide themselves, mimicking conscious choice in their unplanned and to all intents and purposes spontaneous actions, is apparent without knowing anything about their design. You can tell by the way they move that they are free entities.
Looked at coldly, these devices are just inflated plastic balls whose movements are guided by rotors, like a toy drone. Their behaviour is by turns entrancing and mildly menacing. They rise one after another from their resting positions in an upper gallery and calmly hover out into the open domed arena where their human guests are waiting. They are never at rest. As they glide in formation one or another is always changing its position, approaching the people below with what seems like curiosity. Then they all follow. It is when the entire swarm gathers directly above you that it suddenly becomes a threatening, sinister presence.
This artwork that opens visions of a future in which life evolves beyond biology itself.
The true secret of copying life, this installation shows, lies in movement. Dance, the oldest human art, turns out to be a key to comprehending life itself, and reproducing it. The orbs dance with you. They locate and follow members of the audience, not with mechanical inevitability but a complex, gracious harmony. Making and breaking patterns, coming together and loosely floating apart, they dance with each other, too.
Data-driven AI technologies are well suited to address chronic inefficiencies in health markets, potentially lowering costs by hundreds of billions of dollars, while simultaneously reducing the time burden on physicians.
These technologies can be leveraged to capture the massive volume of data that describes a patient’s past and present state, project potential future states, analyze that data in real time, assist in reasoning about the best way to achieve patient and physician goals, and provide both patient and physician constant real-time support. Only AI can fulfill such a mission. There is no other solution.
Technologist and investor Vinod Khosla posited that 80 percent of what human physicians currently do will soon be done instead by technology, allowing physicians to focus their time on the really important elements of patient physician interaction.
Within five years, the healthcare sector has the potential to undergo a complete metamorphosis courtesy of breakthrough AI technologies. Here are just a few examples:
1. Physicians will practice with AI virtual assistants (using, for example, software tools similar to Apple’s Siri, but specialized to the specific healthcare application).
2. Physicians with AI virtual assistants will be able to treat 5X – 10X as many patients with chronic illnesses as they do today, with better outcomes than in the past.
Patients will have a constant “friend” providing a digital health conscience to advise, support, and even encourage them to make healthy choices and pursue a healthy lifestyle.
3. AI virtual assistants will support both patients and healthy individuals in health maintenance with ongoing and real-time intelligent advice.
Our greatest opportunity for AI-enhancement in the sector is keeping people healthy, rather than waiting to treat them when they are sick. AI virtual assistants will be able to acquire deep knowledge of diet, exercise, medications, emotional and mental state, and more.
4. Medical devices previously only available in hospitals will be available in the home, enabling much more precise and timely monitoring and leading to a healthier population.
5. Affordable new tools for diagnosis and treatment of illnesses will emerge based on data collected from extant and widely adopted digital devices such as smartphones.
6. Robotics and in-home AI systems will assist patients with independent living.
But don’t be misled — the best metaphor is that they are learning like humans learn and that they are in their infancy, just starting to crawl. Healthcare AI virtual assistants will soon be able to walk, and then run.
Many of today’s familiar AI engines, personified in Siri, Cortana, Alexa, Google Assistant or any of the hundreds of “intelligent chatbots,” are still immature and their capabilities are highly limited. Within the next few years they will be conversational, they will learn from the user, they will maintain context, and they will provide proactive assistance, just to name a few of their emerging capabilities.
And with these capabilities applied in the health sector, they will enable us to keep millions of citizens healthier, give physicians the support and time they need to practice, and save trillions of dollars in healthcare costs.Welcome to the age of AI.
One of the pioneers in this space has been Australia’s MoodGYM, first launched in 2001. It now has over 1 million users around the world and has been the subject of over two dozen randomized clinical research trials showing that this inexpensive (or free!) intervention can work wonders on depression, for those who can stick with it. And online therapy has been available since 1996.
TAO Connect — the TAO stands for “therapist assisted online” — is something a little different than MoodGYM. Instead of simply walking a user through a serious of psychoeducational modules (which vary in their interactivity and information presentation), it uses multiple modalities and machine learning (a form of artificial intelligence) to try and help more effectively teach the techniques that can keep anxiety at bay for the rest of your life. It can be used for anxiety, depression, stress, and pain management, and can help a person with relationship problems and learning greater resiliency in dealing with stress.
TAO Connect is based on the Stepped Care model of treatment delivery, offering more intensive and more of a variety of treatment options depending upon the severity of mental illness a person presents with. It is a model used elsewhere in the world, but has traditionally not been used as often in the U.S. (except in resource-constrained clinics, like university counseling centers).
Today, TAO Connect is only available through a therapist whose practice subscribes to the service.