There is no need to fear artificial intelligence, yet
The most dangerous beings are still those who handle machinesUpdated: Oct 18, 2019 19:03 IST
The fear of artificial intelligence (AI) is not new. The first time that the world worried that a human-made, and yet superior-to-humans intelligence, has been created and had the capacity to destroy the world was in 1946, when the Electronic Numerical Integrator and Calculator (ENIAC) was built. It was a precursor to computers and could complete in 30 seconds calculations that would take a human being 20 hours. At that time, press coverage on the ENIAC was as alarmist as it has been in the past few years on AI. Some reports called it a mathematical Frankenstein, even a controller of weather. It was, as we well know, none of those things.
The hype around AI now is similar. Articles foretelling the end of times where some form of AI has made yet another human profession obsolete are popular reading. It’s a fascinating thought that humans can create machines that can think and make decisions like us. From this, it’s a small jump to their learning to think like humans. They might then behave like humans too and want to take over the world, because that’s apparently what all humans eventually want to do. But there is no need yet to prepare for an AI apocalypse . Essentially, an AI is a computer programme that can do a specific task that a human has written it for. The thing that sets AI apart from other computer programmes is that the code allows the programme to sometimes rework itself — that it can learn. An AI can also handle far more data than any human and do it several million times faster. Recognising patterns in cell samples, for instance, in order to better detect cancer, is something an AI can do much better than humans. Feed it enough speech samples, and it begins to figure out what humans mean when they make specific noises. Feed it millions of faces, and it can go through all of the faces seen on all of the cameras and identify everyone — facial recognition.
But so far, AI has shown no signs of wanting to take over the world, or even of enough intelligence to warrant any panic. Just as with other machines, the most dangerous beings in the world continue to remain those who handle these machines, humans. Humans, either through ignorance or design, have created machines that can be used to control or kill millions of people — nuclear power, guns and bombs, industrial pollution. AI remains something that is, in the words of Douglas Adams, “Almost, but not quite, entirely unlike” intelligence.