More Than Meets the AI

In which we talk about the liar’s dividend, deepfakes, truthfulness, and how to survive this new reality without losing decency.

‘Autobots wage their battle to destroy the evil forces of the Decepticons.’

In last Sunday’s episode of Last Week Tonight—well worth watching—John Oliver dove into AI Slop, the growing flood of AI-generated content online. He laid out a few solid foundations for the conversation: this new wave is modestly profitable for creators, incredibly lucrative for platforms, and highly corrosive for our perception of reality.

Two moments from the episode stood out to me. First, Oliver echoed something I already heard (maybe my wife?) some time ago: Pinterest is ruined. A platform once built on creativity and inspiration is now a chaotic feed of uncanny, uninspiring visuals. It’s like asking Avatar to help you design your grandma’s backyard.

The second point—more disturbing and more critical—is the concept of the liar’s dividend.

Legal scholars Bobby Chesney and Danielle Citron define the term as a twisted consequence of deepfake awareness. As the public becomes more conscious of how convincing AI-generated content can be, liars become more believable. The more people learn that deepfakes exist, the more plausible it becomes to dismiss objective evidence as fake. How do you trust your eyes when you know your eyes can be fooled—and often are?

It vaguely reminds me of the Epimenides paradox—the Cretan who said all Cretans are liars—but only after getting stuck inside Plato’s cave.

Two examples cited in the article I mentioned earlier are especially telling. One of the January 6 rioters was accused of carrying a handgun. His lawyer claimed the video evidence was a deepfake. Meanwhile, Tesla’s attorneys argued that some of Elon Musk’s past comments about autopilot safety shouldn’t be admissible in court because they might have been AI-generated.

The article also shows that, in Spain, we had early previews of this tactic. Back in 2017, Foreign Minister Alfonso Dastis claimed that many of the photos showing police violence during the Catalonia referendum were fake, despite over a thousand reported injuries. And just as I write these lines—on the afternoon of June 23, 2025—I’ve read that former minister José Luis Ábalos has hinted that the incriminating audio recordings made of him may have been “altered or manipulated.” Days earlier, Santos Cerdán said something similar.

Chasing Truths

Truth is hard work. It is so hard that we Spanish have a word—veracidad—which refers specifically to the pursuit of truth, acknowledging how elusive the goal is. Journalists aren’t expected to deliver the truth (because that’s impossible), but they are expected to chase it.

When the public starts to feel that truth is no longer being pursued with enough conviction—that journalism is tainted by lies, influence, or traffic-hungry cynicism—then the presumption of honesty collapses. Trust erodes fast, and it’s almost impossible to rebuild.

AI hasn’t broken journalism. But it has added another drop—possibly acid—into a glass that was already full. Because this time, the doubt isn’t just about what was reported. It’s about whether what we’re seeing even exists.

You may have heard of the Gell-Mann Amnesia effect, but if not, let Michael Crichton explain:

“Briefly stated, the Gell-Mann Amnesia effect is as follows. You open the newspaper to an article on some subject you know well... You read the article and see the journalist has absolutely no understanding of either the facts or the issues. Often, the article is so wrong it actually presents the story backward—reversing cause and effect. I call these the ‘wet streets cause rain’ stories. The paper’s full of them.”

Now throw AI into that mix and tell me what happens.

A Very Generous Dividend

Here’s the thing about the liar’s dividend: you don’t even need deepfakes. A criminal can record a conversation with a friend and accuse a third person of wrongdoing. Whether the person did it or not doesn’t matter—headlines still carry their name. Even if it’s your enemy, record yourself blaming them and let the media and the algorithm do the rest.

As El País journalist José Manuel Romero reminded his readers, Judge Javier Gómez Bermúdez laid out in his verdict on the 2004 Madrid train bombings how the campaign to falsely accuse ETA worked:

“One detail is isolated, removed from context, and presented as if it alone justifies a conclusion—while ignoring the need to evaluate evidence as a whole using logic and experience.”

Sound familiar?

In Spain, legal defense strategies often involve a version of this: turn on the shit-sprinkler, flood the conversation with distractions, and hope that by generating enough isolated or decontextualized claims, the original accusation drowns. Constant leaks from investigations mean that anyone mentioned in any report suffers a kind of media conviction that tarnishes their reputation, especially their digital footprint, whether or not they’re ever cleared.

Here’s the paradox: we usually accept that conviction when we dislike the person in question, and reject it when we support them. In times of hyper-polarization, that only makes the problem worse. Public punishment ends up being carried out by whichever side already hates you. And if you didn’t start with the benefit of the doubt, good luck surviving it.

Take Dani Alves. Half the Spanish public believes he’s a rapist. The other half thinks he’s innocent. Contradictory rulings have only deepened the divide.

Reality is stubborn. And messy. I’ve often said I might be one of the few men who could survive a full audit of my WhatsApp by my wife—yes, I only have one phone. But surviving an audit by someone who wants to destroy you? That’s another story.

Sarcasm, jokes, hedging, trying to placate someone to avoid conflict… Who hasn’t let something slide in a group chat to keep the peace? But if someone wants to take you down, the headline is ready: “X stays silent as friends make indefensible comments.”

We can quote all the Latin we want about Caesar’s wife needing to appear virtuous and be virtuous. But let’s be honest: after the whole knife business, Caesar’s wife was pretty much screwed.

Trap for Fools

Trump-style populism has poisoned our relationship with reality. A life of ethical, meaningful work can go up in smoke if a media outlet wants a particular number of clicks.

Richard Jewell helped prevent a terrorist attack and was crucified by the press for living with his mother and being overweight. Meanwhile, a swarm of crooks, sycophants, and incompetents can take part in the administration of the most powerful country on earth and not only avoid punishment, but thrive.

So, what can we do?

Being a good person is always a start, but it’s no guarantee. Avoiding conflict altogether? That’s just letting the worst people win. Developing a rigorous personal filter and only trusting those who’ve earned it? It’s noble, but elitist by definition. Reading books like Bulos: manual de combate by Rubén Sánchez? Sure, that helps—but how many people have the time or energy to do what he’s been doing for years?

Life isn’t fair, and maybe we’re at a moment when it’s worth remembering Rudyard Kipling:

"If you can bear to hear the truth you’ve spoken
Twisted by knaves to make a trap for fools,
Or watch the things you gave your life to, broken,
And stoop and build ’em up with worn-out tools."

Ultimately, maybe the best we can do is fasten our seat belts and wait for a better social climate, contribute honestly to the debate, resist unjustified polarization, and keep cynicism at bay. If you’re lucky, rely on the moral compass your parents gave you—or the one you’ve built over time.

Despite being an atheist, I try to follow the eighth commandment: “Thou shalt not bear false witness against thy neighbor.”

And if you’re unsure whether a witness is false, sometimes silence is the only decent answer.