Artificial intelligence still has a long way to go, says IBM’s head of AI, Alexander Gray
“The biggest misconception is that we have it. I wouldn’t even call it AI. I would say it’s right to call the field AI, we’re pursuing AI, but we don’t have it yet,” he said.
Gray says at present, humans are still sorely needed.
“No matter how you look at it, there’s a lot of handcrafting [involved]. We have ever increasingly powerful tools but we haven’t made the leap yet,”
According to Gray, we’re only seeing “human-level performance” for narrowly defined tasks. Most machine learning-based algorithms have to analyze thousands of examples, and haven’t achieved the idea of one-shot or few-shot learnings.
“Once you go slightly beyond that data set and it looks different. Humans win. There will always be things that humans can do that AI can’t do. You still need human data scientists to do the data preparation part — lots of blood and guts stuff that requires open domain knowledge about the world,” he said.
Artificial intelligence is perhaps the most hyped yet misunderstood field of study today.
Gray said while we may not be experiencing the full effects of AI yet, it’s going to happen a lot faster than we think — and that’s where the fear comes in.
“Everything moves on an exponential curve. I really do believe that we will start to see entire classes of jobs getting impacted.
My fear is that we won’t have the social structures and agreements on what we should do to keep pace with that. I’m not sure if that makes me optimistic or pessimistic.”