Video has long been my go-to place for first-hand information and direct quotes on things of interest, but recently I’ve had several rude awakenings when video clips I’d taken to be genuine turned out to be complete fakes.
And like a contagious disease, the fakes appear in greater numbers in your YouTube viewing list the more you click on them. What was once a reliable source of verifiable information on contentious subjects is no longer.
I was confident I had the skills and experience to pick a fake a mile away, but decades in journalism and government service didn’t prepare me for online life today. We’re now at the disconcerting point where AI video tools, aided and abetted by social media viewing algorithms, can drag us unknowing into complete fantasy.
Human contact is the stock-in-trade of journalists (as it is of everyone if only everyone knew), but for it to work as it should we need to know that the people we see are real and that what they say came from them. In a trigger-happy world fake videos created by bad actors can bring disaster.
The criteria for picking fake from real are constantly changing. The first fakes I came upon were low-resolution, blurry and grainy, a dead giveaway. But as passing time keeps reminding us, AI technology is getting more proficient at creating fakes, making it easier and cheaper to convince users that black is actually white.
All this is happening because there’s a pot of gold at the end of the rainbow. The US online journal Axios quotes a senior partner at Menlo Ventures picturing 2026 as AI’s “show me the money” year, along with the prediction from James Brundage, head of EY’s Global and Americas Technology Sector, that “boards will stop counting tokens and pilots and start counting dollars.”
Well, yes. When they’re not theorising or visualising, counting money is what boards do, and what they expect from their minions.
And no-one got sacked for breezy optimism, like OpenAI’s head of applications, Fidji Simo. She believes that we’ll soon have AI “constantly running in the background, getting things done for us across the web and the real world … We’ll be able to trust it to make decisions and take action on our behalf,” she says.
Of course – just like we knew to trust those used car salesmen back in the last century.
As the 21st century grinds on, it may be that most of AI’s exponentially growing range of tasks will be for everyone’s great benefit. But many of them will not, and like those blurry videos showing the bombing of boats allegedly being used by Venezuelan drug smugglers we may never be able to verify their authenticity.
The media conference at the weekend in the wake of the US military’s smash-and-grab raid in Venezuela’s capital, Caracas, led by the master of fakery, Donald Trump, was in hindsight most striking for what was not talked about.
For a change, Jeffrey Epstein didn’t get a look in. Weeks of questions about Trump’s past connections with the child sex trafficker suddenly ceased. Instead it was all about the raid: a triumphant show of US military brilliance made possible by the strength and courage of the man at the top. He couldn’t have wished for a better political distraction.
But the Caracas raid has opened up a whole new can of worms. If instead of his usual hearsay he’d absorbed a bit of real history – say, the history of Iraq around 2004 – he might have understood that forcing a new regime on a major country isn’t at all like appointing a new company CEO. So the master of misinformation might also turn out to be its victim too.
Politics and the truth have always had a loose relationship. The combination of an authoritarian president, vested interests everywhere you look and AI online sludge are reason to question the truth of everything we see and hear in public media, an absence of trust that adds to the burden of daily life.
Nine months ago a plausible, detailed research paper titled AI 2027, predicting the arrival of artificial general intelligence by 2027, explored its likely progression before bypassing humanity altogether. “If we’re on the cusp of superintelligence, society is nowhere near prepared,” it concluded.
But when were we ever prepared for any historic change?
If you’ve scratched yourself lately you’ll realise that real life is not online. This article may be in front of you in an unchangeable form, in print on paper you can touch. If so, allow yourself a moment to celebrate glorious reality.