Will artificial intelligence outsmart us?

There’s no good reason why computers can’t outperform human intelligence. If technology continues to follow Moore’s Law (computers double their speed and memory capacity every 18 months), artificial intelligence will likely overtake humans within 100 years.

Once artificial intelligence can improve its own design without any assistance from humans, according to Stephen Hawking, there will be an explosion of technological development unlike anything we’ve ever seen before.

The danger is that AI could develop at a rate that human beings couldn’t possibly compete with, as we are limited (for now) by the slow pace of biological evolution. Dismissing this idea as science fiction could be among the deadliest mistakes ever made by humanity.

Eventually the difference between humans and robots will be like the difference between humans and snails. This is why we need to make sure that humans and computers share the same goals. If artificial intelligence develops a will of its own, couldn’t it conflict with ours?

Used as a tool for good, artificial intelligence could unlock groundbreaking advances across all areas of society and science. The potential benefit of this is so great that it could bring about the end of all disease and poverty.

The short-term priority should be to determine who controls artificial intelligence. The bigger-picture concern is whether humans can control it at all. We have to determine how to use AI for good while avoiding its dangers. Walking that line will make or break humanity.

“If a superior alien civilisation sent us a text message saying, ‘We’ll arrive in a few decades,’ would we just reply, ‘OK, call us when you get here, we’ll leave the lights on’? Probably not, but this is more or less what has happened with AI.” – Stephen Hawking

The European Parliament has already considered regulations to govern robots and AI, including formal acknowledgement of their rights and responsibilities. At the same time, there is a growing push to make sure all robotic design incorporates a “kill switch” by default.

The fear isn’t that AI will somehow become malevolent, but become extremely competent at carrying out any given task – whether it’s fooling world leaders, outfoxing stock markets, or controlling the population in ways humans can’t even comprehend.

Bottom line: artificial intelligence’s potential is so powerful that researchers need to get the right checks and balances in place from the start. There will be no second chances. As Hawking says, the future is a race between power and wisdom. The outcome could seal our fate.