Stephen Hawking has given another warning about artificial intelligence (AI), but there is good news too.
Stephen Hawking has warned about AI before. But he has done it again, speaking at the opening of Cambridge Universities’ new artificial intelligence centre.
“The rise of powerful AI will be either the best or the worst thing ever to happen to humanity,” he said, but added: “We do not know which.”
He continued: “It will bring great disruption to our economy, and in the future AI could develop a will of its own that is in conflict with ours.”
There are some issues relating to AI that we may overlook.
For one thing, bear in mind that our species, Homo sapiens, is around 10,000 generations old. We are not old. Our brains have evolved, and while we roughly understand the mechanism by which evolution works, no one knows how our brain developed the way it did, or indeed why, – not for sure. We do not understand how the mind works. How can we be certain that AI will not evolve in a comparable way? In a digital environment, an evolutionary process might see generations separated by no more than nano seconds. Ten thousand generations of a digital technology could evolve in less than an hour.
No one can say for sure that a process, which no one understands, cannot create an approximation to something we do not understand.
But at least we are no longer ignoring the danger, unlike a couple of years ago when warnings about AI were seen as daft, and such arguments were dismissed, suggesting their authors watched too much science fiction.
In the US, Elon Musk has been ploughing money into a think tank designed to ensure that AI is developed as a force for good.
The very same artificial intelligence centre, whose launch Professor Hawking was speaking at, is partly charged with trying to understand the dangers of AI.
In the US, Barack Obama has spoken on this subject.
It is being take seriously, at last. That is good news, but frankly this is perhaps the single most important topic of this age.
But whatever happens, AI will change us, and do so fast.
It may enhance us, via augmented reality and even via interfaces directly to our brains, perhaps using neural lace. Maybe we will able to download modules for understanding a tricky concept. Want to speak Mandarin? Download this software into your brain. Not that you will need to speak Mandarin when we all carry real time voice translators with us – technology that may be less than a decade away.
Let’s say Homo sapiens is exactly 10,000 generations old. The next generation, number 10,001, maybe so changed by AI that they are effectively a new species.
AI, good or bad, our species may be a generation away from becoming Homo sapiens artificialis intelligentia. Many of us may live to see it. The decade after next, may even see it.