That fine came under General Data Protection Regulation (GDPR), but while regulation has somewhat put guardrails on how tech companies use that, there are seemingly fewer on the use of copyrighted material. And the tech CEOs have such a seemingly loose handle on the idea of copyright that this is becoming increasingly important, not just to media groups but individual writers and academics.

The AI revolution — the rambling, multidirectional thing that it is — has come about because of technology companies’ scraping of copyrighted art and text, gathered from all across the internet.

It would be churlish to say AI is without uses, such as the automation or improvement of repetitive tasks, analysing medical data, or simplifying industrial processes.

But it’s easy to get ahead when a large amount of your foundational material isn’t being paid for.

While OpenAI and others have signed some licensing deals with publishers around the world, both news and academic, this is a bit like shutting the door after the horse has bolted. Or, perhaps, more like actually fitting a door onto the barn, then shutting it.

There has been a cavalier brashness from tech companies, even by their standards. Meta boss Mark Zuckerberg, riding a wave of self-reinvention powered by AI interest after his failed pivot to the metaverse, told The Verge earlier this year that “individual creators or publishers tend to overestimate the value of their specific content in the grand scheme of this”.

The plethora of lawsuits against the companies for copyright infringement — News Corp against Perplexity AI, The New York Times and others against Microsoft and OpenAI, the produce of Blade Runner 2049 against Tesla over AI imagery — suggests otherwise.

That Zuckerberg’s company is now suing the DPC, which has fined it over data breaches in the past, smacks of trying to have its cake and eat it too.

This week, Microsoft chief executive Satya Nadella pleaded for copyright laws to be loosened so it’s easier for AI companies — and Microsoft is one of the biggest players, given its huge stake in OpenAI, its own development division, and the prevalence of its software worldwide — to train on copyrighted data. He wants a widely-agreed definition of “fair use”. This would, he doesn’t say but we can all understand, mean the company doesn’t have to shell out for licensing.

“What’s copyright?” Nadella asked at an event in London. “If everything is just copyright then I shouldn’t be reading textbooks and learning because that would be copyright infringement.”

This seems a wilfully rudimentary understanding of the issue by an intelligent man. Even school students are taught that they have to cite their sources when coming up with arguments or essays; generative AI has no such obligation, and asking it to cite sources for the information it produces mightn’t even give a user the right answer to begin with.

So many billions have been thrown at the industry, it’s unclear if it will ever make the money back. OpenAI alone is on course to lose €5bn this year yet it’s still getting investors. An Irish SME making losses amounting to OpenAI’s spare change wouldn’t get that sort of leeway.

Similarly, if one publisher were to take the material of another and use it in their own publication without permission, they’d likely get an invoice at the very least — as it would were a publisher to use an article by a freelance writer.

One wonders how happy Microsoft would be if it found another company taking its software and incorporating it into another app. Not very, one imagines.

If that can be so easily understood, then sure the concept of paying for usage (annual Office365 subscriptions anyone?) isn’t beyond Big Tech.

Surely it’s not too much to ask that some regulatory body gets the powers to make sure the biggest companies in the world pay their way as well.