Sourcing Truth: A Future of Synthetic Information
Falling down the rabbit hole on the hunt for accuracy
Earlier this week, ZINE published Julia Dixon’s 3_TRENDS volume, an interview series, uncovering overlooked social shifts.
In our analysis of Neo-Luddism, we shared:
“Despite the pushback [on AI], advancements aren’t going away, with as much as 90% of online content predicted to be synthetically generated by 2026.”
After publishing, one reader reached out questioning the audacity of this figure. I also think it’s worthy of a follow up...
I thought the stat was hyperbolic, and considered this forecast and its likely reaction to be fitting for our trend around unchecked AI hype and “anti-tech” sentiment.
All “predictions” require a sense of skepticism — after all the future does not yet exist.
The “90%” was sourced via Futurism.com which reported the figure from the Innovation Lab of Europol, the European law enforcement group. Not a shabby source, I thought.
But after my piece was out and I read it back live on the internet, I wanted to go deeper into this “90%.”
Maybe it was too audacious...
As Will Cady puts it,
“Phones used to distract. Now they distort.”
This distinction is paramount, and the latter is more prevalent yet less perceptible. We require diligence.
Down the reference rabbit hole…
If you currently click into Futurism’s source link, navigating to the Europol report, Facing Reality: Law Enforcement and the Challenge of Deepfakes, and then Command-F search for any mention of “predict,” "estimate," "90," "ninety," "percent," or "2026," you will not find the claim: “as much as 90% of online content is predicted to be synthetically generated by 2026.”
It’s not there.
Yet if you Google "90 percent of online content synthetically generated 2026" you’ll find sources like Psychology Today, Breitbart, Vox and eMarketer all using the 90% in 2026 claim. They all source Europol.
But again, it’s not in the report.
After reaching out to the author of the Futurism piece, which was one of the first publications to cite the “90%,” she shared that Europol actually updated their report recently. That’s why the “90%” can’t be found in the current version online.
By using The Internet Archive’s Wayback Machine, we can find a version of the report which did include the claim.
Phew. We’re not crazy.
There it was:
“Experts estimate that as much as 90%⁴ of online content may be synthetically generated by 2026.”
Thankfully, all these publications didn’t mindlessly source one another and a stat that didn’t exist. It was in fact published by Europol.
Catch that, though?
There’s a superscript ⁴. The “90%” had a footnote.
Europol’s Innovation Lab did not produce this figure. It was sourced.
⁴ Schick, Nina, Deepfakes: The Coming Infocalypse: What You Urgently Need To Know, Twelve, Hachette UK, 2020.
Nina Schick is a political commentator, advisor, and public speaker, specializing in how technology is reshaping politics in the 21st century. She’s a regular contributor to international broadcasters including CNN, Bloomberg, Sky, and the BBC, and her work has been published by the Times, the Telegraph, and the New Statesman.
Schick is incredibly respected and a credible source. Her book and bio assuaged my concerns even if the “90%” in the report was just from desk research.
But at this point, I was too far gone.
In reading Schick’s 2020 work Deepfakes: The Coming Infocalypse: What You Urgently Need To Know, you’ll come across a passage on page 47. This is where the “90%” originates from.
Schick writes,
“[...] I spoke to Victor Riparbelli, the CEO and Founder of Synthesia, a London-based startup that is generating synthetic media for commercial use. According to Victor, video is by far the best way to deliver information, and synthetic media is a ‘glimpse into the future of how humans will create content.’
...
As the technology improves, Victor tells me, it will become ubiquitous. He believes that synthetic video may account for up to 90 per cent of all video content in as little as three to five years. This is a great thing, he says.”
And there we have it.
The claim “as much as 90% of online content is predicted to be synthetically generated by 2026” does not in fact originate from Europol nor Schick herself, but from Victor Riparbelli, the co-founder of Synthesia, the “#1 AI Video Generation Platform.”
What happens after we find truth?
I’m conflicted.
First, I feel a responsibility for amplifying such an absurd number without the explicit asterisk that I personally considered it so absurd in the first place. As I learned of its true origin, without a forewarning, re-publishing “90%” dangerously affirms a legitimacy, which it does not deserve.
Second, the rigor to de-bunk every claim on the internet is insurmountable — an impossibility. We require trust. But this trust is hard-earned and easily tarnished. What’s required to verify even the most obviously false claims takes hours. This is worrying in its own right, but when we consider that our platforms want more content produced, and more frequently, accuracy is decentivized. In fact, even if one wanted to clarify a published inaccuracy, reaching those exposed to the original message is not guaranteed and if anything, incredibly unlikely.
Third, perhaps our worries around AI-generated images, videos and voices are misplaced. We’re clearly struggling with non-AI produced misinformation and exaggerations on a daily basis. Publications’ sourcing infrastructure and diligence for accuracy is table stakes — especially from Europol no less. Perhaps Europol also went down the same rabbit hole and this is why we can no longer find “90%” in their report’s current version.
And lastly, maybe Riparbelli is actually right here.
Maybe Riparbelli’s 90% forecast is absolutely accurate. This story is case in point. A shit load of information on the internet is utterly synthetic. Completely fabricated. Nonsense. Made up. Lies.
Our current, collective hysteria asks, “What happens in a future when we can’t tell if a video is real or not?”
Not only is fictitious information already all around us, but we don’t even seem to mind.