We can’t stop the development of Artificial Intelligence by government decree any more than we could have stopped Covid by government decree.
The sum of human knowledge used to be an unimaginably vast entity, consisting of acres of books and images, stretching across countless libraries, that could never be known by one generation, still less by one human. Now, thanks to AI, it can exist all in one place, a server farm in California, along with the algorithms that bombard it every second of every day.
The AI cat is out of the bag – $350 billion has already been invested in commercial applications, from internet search to composing music to gold prospecting.
Who knows how many tens of billions governments have already spent on military applications of AI each year? That’s a secret. The next global war will probably be fought by computers. The next generation of drones is probably being field-tested in Ukraine as we speak.
AI is potentially as big a threat to humanity as Covid was, or bigger. And many are now calling for tighter controls. But a recent appeal by Yuval Harari and several hundred leading experts, for AI research to be halted or slowed immediately has had no effect, and it may be in years before any such edict could possibly be agreed and issued, by which time according to Harari himself, AI bots may have moved beyond human control.
Harari fears computers which assimilate and recombine all human knowledge far more quickly and efficiently than humans, could take control of that knowledge unimaginably fast. He argued in a recent New York Times article, that AI could control humanity by controlling our language. He is right. All knowledge is language. Even images are a kind of language.
Nobody owns language, and never can. But fortunes are spent on pure language – the advertising, publishing, and computer software industries, have between them $1.3 trillion annual revenues. And some languages are more powerful than others.
At the heart of the debate is copyright. Who owns the entirety of human knowledge when it is recombined in new ways? Does it belong to the OpenAI computer company that is the current market leader? Or does it belong to the countless writers, researchers, publishers, photographers, filmmakers, to name but a few, whose work has now been hoovered up to feed the algorithms? And by far the largest part of what is in the AI/big tech memory banks is the entire history of all our social media, and email. It belongs to us, to all of us. But if we want to retain our ownership, the only way is to start fighting for it immediately.
The key questions is what barriers, if any, should there be to universal access to this entity, “the sum of human knowledge?” If it belongs to all of us, then should we all have equal access to it? And what of the harm it could cause, if AI became an arms race, in the non- military sense, between rich and poor, employer and employee or among and between local communities.
I oppose banning or slowing AI research because there is better approach, straight out of the Covid playbook. Just as the Covid vaccine was developed by an unprecedented global collaboration between scientists, the same is needed to confront AI and ensure that it is used for good – to protect freedom, democracy and the global environment, and to prevent or even eradicate, war and poverty.
That will involve scientists of course, along with engineers, writers, doctors, lawyers, politicians, artists, philosophers, soldiers, marketeers, and many others.
I invite any one interested to contact firstname.lastname@example.org to campaign for a global task force that is independent of major factions such as governments, NGOs, financiers, industry lobbyists, and hustlers of all sorts, far too many of whom are signatories to the letter referred to in the New York Times.