What happens when you say goodbye to fact-checkers on social media?

A series of posts by Meta CEO Mark Zuckerberg on the Threads social media app, outlining his changes to content moderation. PA Wire

Facts, it seems, aren’t what they used to be.

The 20th century American comedian Groucho Marx famously said: “Those are my principles, and if you don’t like them… well, I have others.” It is a sentiment that is said to predate him. But if we jump forward to the 2020s, we might be forgiven for reinterpreting the phrase to be about "facts". It seems that facts are no longer seen as sacrosanct, nor worth preserving as a principle.

Earlier this month, Mark Zuckerberg, the chief executive of tech giant Meta that owns Facebook, Instagram, WhatsApp, among others, said his company is doing away with fact checkers for Facebook and moving to the model of "community notes" used by fellow tech owner Elon Musk at X, formerly Twitter.

I don’t need to rehash the discussions from recent years about how social media has been warped by fake news, misinformation and disinformation. Nor the debates about whether we are now in a post-truth era. Even less about how the echo chambers created by algorithms hungry for engagement drive tribalism – not just online but offline too, and society fracturing into seemingly unbridgeable tribes.

It feels unnecessary to pose the question in a newspaper column, but isn’t it foundational that we accept that some things are not conjecture?

The irony of writing an opinion piece about facts is not lost on me, but when there are no shared facts, or newsworthy reliable sources, what do we do? Do we end up living in a real-world equivalent of shouty people who tell us something is true just because they received a WhatsApp message? Are we to assume that any kind of information we receive is a fact?

Every wave of scientific discovery is founded on facts. So I only pick the Renaissance and Enlightenment to contrast with the recent western announcements that facts are no longer the "in" thing.

There are consequences to such developments. Some of these are of a much more long-term nature and that should concern us all.

It is namely this. If digital inputs that are being increasingly used to develop AI are free of facts, what will that do to the knowledge created, and to its proliferation into every part of our knowledge and societal ecosystems?

If we dilute all the knowledge in the world, a lot will be undone, and the price will be paid by coming generations

The real-world impact of AI’s biases are well documented: everything from wrongful jail convictions based on poor AI facial recognition, to rejections from job interviews because the AI recruiter deems you don’t match a pre-existing notion of someone who can do a particular job satisfactorily.

Artificial intelligence is driving a change in the way knowledge is processed, produced and used in the world around us. It is widely acknowledged that it contains biases – after all its inputs originate from an already biased world. And those biases are exacerbated by the biases of the people building the models, rules and applications.

But what happens if the real world is polluted with information that is not fact? Already we’re not able to deal with the biases that exist in the world. But we are entering a time where the guardrails on facts are being taken down at the very moment that the knowledge being fed into these systems is informing young people and defining their views and beliefs. This could very likely hold true for generations. It is one reason why the flippancy with which facts are treated should be seen as deeply dangerous.

There’s an argument for "offline" real world knowledge, such as that found in books. But that can't spread at the same speed as AI, nor have the reach and implications across the systems that run the world.

I may recall checking encyclopaedias to complete my homework, but for younger people, born surrounded by constant internet and content, access to high quality "knowledge" has been replaced by the need to gather as much information as possible, from what can often be dubious websites.

When internet access and in particular social media emerged, the great hope was that it would be a means to democratise access to knowledge, enable being heard, and add to knowledge from new sources. This democratisation was subverted by algorithms and the opinions which caught their attention and turned those opinions into what are often taken to be facts.

If at this moment we dilute all the knowledge in the world, that has taken centuries to accumulate, a lot will be undone, and the price will be paid by coming generations. Facts matter. Otherwise what will the knowledge legacy be that we leave for our children and grandchildren?