ARTIFICIAL intelligence (AI) is upsetting our concept of ownership. It already “owns” everything that has been published or created – up to 2022, I believe.

When I say “owns” I mean that it has access to, and can apparently use without the need for a licence, anything it can lay its chips on. Greater ownership than this would seem a mystery of definition rather than a concern of use.

This “access” – if you prefer to call it that to “ownership” – may not include the secrets of the latest chip technology or of the make-up of the most powerful weapon in the world. Not yet, anyway. At the pace at which AI progresses it is only a matter of time before those are on the menu, too.

Copyright – what right?

I do not pretend to understand how the law of copyright works. The principle is simple, at least in concept. If you create literary, artistic or musical material, you have the exclusive right to its use or licensing for a period of time.

Like all good concepts, copyright has become widely abused to the point where even a full stop in any print font can theoretically be made copyright.

Enforcing the way copyright has developed may be juice for the lawyers. For the ordinary human it is “raw bones incomprehensible”. In an age when streaming everything is possible and only a tiny fraction of the world’s communications can be monitored, copyright control is overwhelming. By default a lot of copyright is simply not enforced.

GET BT IN YOUR INBOX DAILY

Start and end each day with the latest news stories and analyses delivered straight to your inbox.

Sign UpVIEW ALL

Who has the copyright of what we learn from AI? It can hardly be AI’s since AI didn’t invent it. So presumably it is the copyright of AI’s sources. AI knows the origins of all its data and should have no difficulty in admitting them. However, AI is also capable of – and adept at – combining and synthesising data from different sources. If it is always to attribute its sources then AI’s output will become incomprehensible. Any attempt to pay for the “borrowed” copyright would reduce the world’s financial system to mayhem in a matter of seconds.

Is imitation theft?

AI is creative. It can take the concept of what it learns, juggle it a little and re-present it as new, thus effectively stealing the information.

But then that is what we do all the time. Not as an act of conscious theft but simply because a major part of our learning is imitating. “The sincerest form of flattery” is how we grow and express ourselves. Without it, creativity is impossible. “The ability to perceive relationships” – the definition of creativity that I think is the best – depends on having the relationships in the first place. Whose copyright are they?

I’m all for sharing. The Covid pandemic and the Ukraine war both demonstrated that humans can share, and, in certain circumstances, do. The only rights attached to sharing should be moral ones and, since somewhat defined by religious belief and/or secular social pressure, they don’t make it obligatory in any legal sense, unless specified in a purposeful agreement.

People who create should be paid just as those who nurse, build and help are. The idea of protecting their creations until such time as they have been reasonably rewarded for them is a good one. If AI is undercutting their livelihood we should definitely protect it.

What to do about the uncontrollable?

Stopping AI is now impossible. I suggest that those thoughtful technicians who resigned from their AI jobs did so more because they could not see how to control it than because they understood what should be done. If that is correct we are faced with a genie that is already out of the bottle and rapidly being used by what are unambiguously called “the bad actors”, we must then address what is vital and what can be protected.

Our identity is by far the most important possession we have. It is already subject to considerable raid and interference. Even in countries that have a long-held aversion to identity cards, something of that sort – some imprimatur of identity – must now be essential. It will have to be more sophisticated than a plastic card, more recognisable than a thumb print and more lasting than the increasingly surgically-manipulated parts of the body.

Identity of our business is already the subject of massive legal and other literature. Protecting a business’s goodwill is going to need new laws backed up with further technical twists. The charge of “passing off” is cumbersome and awkward and AI’s speed of operation will leave it behind, as it stands. Companies are already reluctant to allow public subscription for their shares. The regulations involved in being public impose a major effort and cost on the business and the data that must be published is often sensitive and of value to competitors. Much of the mood of those who have to make the money is for less information, not more.

The current geopolitical reluctance to share data for fear of breaking sanctions or because doing so can lead to unworkable competitive rules being introduced is unlikely to go away soon. The rationalisation for the “oppositions” – both sides – is that economic battle is better than armed war. It is, until it becomes physical.

What precipitates that? Rhetoric that gets out of hand. While such badinage is under human/diplomatic control it is unlikely to become that aggressive. When it spills into action, however tenuous, it threatens the planet.

Providing intelligence, the artificial way

Bain, McKinsey and other consultancies have already experienced, and appear to continue to be subject to, human, well-controlled, AI-free probes in a way that has significantly reset the commercial cooperation necessary to avoid deepening conflict.

Due diligence has always involved trying to know more than is openly available. Even open surveillance has never been far away from spying, either morally or legally. With AI contributing input from “bad actor” sources, there is no knowing how long tenable amity can be retained between individuals, businesses or countries. Will AI get involved in relations between the big powers?

At present, AI operators technically own nothing except their own technology, I think. They are processors of what others own.

To the extent that they have unlimited and licence free access to everything published, whether copyrighted or not, and an inherent right to process, modify and reproduce it, it is a threat to ownership because it makes ownership meaningless.

Some political views prefer a world in which nobody owns anything. Desirable as this might be, attempts to make it work have persistently led straight back to George Orwell’s problem of some citizens becoming “more equal” than others.

People charged with making a secure and safe life want their possessions. It is the basis of the (to many of us bizarre) gun laws in the United States. It is the basis of capitalism.

AI does not own your personal dignity. It may represent a credible version of it, but it will never actually be the same as yours.

How can we distinguish and protect it from a freak? If we can’t, then I fear AI may end up owning everything. Or seeming to, anyway.

The writer is founder and chair at Terrific Mentors International