News media versus AI: What if we win?
LIKE all good foundations, the ones on which the media industry relies are usually invisible. We only see what is built on top — a flourishing, plural creative sector. Until, that is, it starts to collapse. The media’s foundation is copyright. As dry and dull as concrete. But if you take the concrete away, a lot of buildings fall down. It is the same with copyright and journalism.
We know this because we have been living it. The internet has undermined our foundations, messing up our relationship with our audience and crippling our business models.
Diminished and enfeebled over the last decades, things are only getting worse for the news industry.
When copyright was undermined, we did not feel the effects immediately. By the time we did, it was already too late.
Generative AI is widely predicted to accelerate the destruction, and it certainly has the potential to do so.
The remains of the media could easily dissolve and drown in the AIgenerated flood of garbage which is already flooding the internet and threatens to wash away what is left of our businesses.
But, unexpectedly, the opposite could also happen. AI might be about to give us back a solid base on which to rebuild our industry and spark a new creative renaissance.
One of our key foundations might be shored up and made solid again.
Copyright was stolen from news business at start of online age
Copyright is a gloriously simple thing. Anyone who wants to copy work you created must have your permission. That is basically it. There are few, and mostly quite specific, exceptions to this general rule.
The problem is that when the internet was too new for the consequences to be obvious, we made the big internet platforms one of those exceptions.
Laws like the Digital Millennium Copyright Act and the notorious section 230 of the Communications Decency Act in the United States, the e-commerce directive in the European Union and others around the world created special legal immunities and exceptions which shielded internet businesses from many legal risks.
As a result, any content created by our industry, the result of our expertise and hard work, became fair game for their use. Everyone’s work was made theirs for the taking.
And they took it. This cost-free, risk-free, access to a nearly unlimited source of raw material, unhindered by copyright and other laws, stole the foundation of the media industry and handed platforms like Google and Facebook an extremely profitable one of their own.
They were able to build trilliondollar businesses on the back of other peoples’ efforts.
The plain fact of it is that they would not be worth anything without free use of “content” created by others.
I think, perhaps unfashionably, that generative AI might be about to change all that.
AI companies regard everything on the open web as fair game for them to use.
Of course they do. If it is OK for search engines, they say, it is OK for them. They have been using everything they can find, in private at first, and now very openly, to “train” their AI systems. They take full advantage of everyone’s work being available to them for free.
But what if it was not?
Marc Andreessen, the quintessential Silicon Valley investor, knows the answer: “Imposing the cost of actual or potential copyright liability on the creators of AI models will either kill or significantly hamper their development.”
Worry less about over-hyped AI start-ups and more about world’s creators. It would cost him a lot of money, in other words. But, I would argue, it would produce much better outcomes for everyone else.
Sam Altman says that OpenAI would not exist if it had to get consent from creators. The news industry also cannot exist if AI continues to use its work to create a competitor.
The risk that the current generation of AI might not be viable is muttered darkly, as if it is a dire threat to the world. It should not be heard as such. If an early version of a new technology turns out not to be viable, innovators can return to their labs and try to improve it. The next version will be better.
If copyright poses a threat, it is not to AI per se, but to over-hyped, overvalued and over-funded AI companies. The world should worry less about them than the wider interests of creators, societies and the creative economy.
The argument is being played out right now. AI companies assert everything they do is allowed under the US defence of “fair use”. Others disagree and the law elsewhere in the world also stands in their way.
The arguments are wending their way through courts and legislatures all over the world. AI companies will need to win all of them to get away scot-free. That is a tall order, in my view.
If AI companies somehow win on this, we will remain in the online world we inhabit right now, managing our ongoing decline and watching as the internet fills with AI-generated grey goo.
What if news media wins copyright battle with AI companies?
But if they lose — and I think and hope that copyright will win — the internet will become a very different place. A better place. It will seem quite small, at first.
AI companies (and many others) will have to ask permission before they copy and use content they find online, just as copyright law demands. If they use stuff without permission, they will be running a gigantic legal risk.
The AI sector, and the activities of anyone else who is taking and using without asking, will need a reboot. Deals will need to be done and maybe some AIs will have to go back to the drawing board.
The foundations of our industry, of the whole creative sector, will be re-solidified. The scope of fair use and legal exemptions will be narrowed. Not asking permission will be a bigger risk for tech innovators. The real role of creators and creative businesses, as the apex creators of value on the internet, will be recognised at last.
But are we ready? Not really.
This process of seeking and giving permission, simple as it sounds, has been left behind by technology.
Read more on www.newsday.co.zw