People are Beginning to Understand the Limits of AI

It has been predicted that AI will add $16trn to the global economy by 2030. Sundar Pichai, Google’s CEO has described AI technology as more important than fire or electricity in terms of the technological development of humanity. It has changed the way we do certain tasks dramatically. We now use AI to recognise songs, draft email replies, use facial recognition technology to unlock our phones and even police borders. A lot of people fear the huge change that AI is predicted to bring, as it has been said that these clever computers will take jobs away from humans and create a mass wave of unemployment.

But recently, people have begun to doubt whether today’s AI technology will actually change the world that much. AI has its limits, and we’re beginning to understand them.

In spite of the success of AI, the fact remains that the hopes and dreams we had about the ways it could be applied to have not materialised. Self-driving cars, for example, are nowhere near being safe enough to be put onto the streets yet. Because of this stagnation, and lack of progress, enthusiasm for AI is simmering down.

So why is this happening? Well, although modern AI techniques are powerful, they are also limited, and can be difficult to implement for a number of reasons. Firstly, AI is expensive. But even if it weren’t, the data that is used to teach these machines how to do things comes from humans. This is a problem because just as AI learns how to complete tasks by observing hundreds of thousands of examples from humans, it also learns the mistakes that we make, without learning from them. Put simply, this means that they are not “intelligent” in the way that people think they are. They are pattern-recognition tools, but they do not have cognitive abilities that organic brains have. They generalise from the rules and patterns they discover in data, and they do not use reasoning.

In short, we don’t know what AI will go on to achieve in the future, but it’s becoming clear that it won’t create driverless cars or add $16trn to the economy any time soon.