AI Video Killed the Truth

As journalism accelerates toward an "extinction-level event," AI video is creating an "Oppenheimer moment" in the war on truth.

I woke up this morning to the latest "Elon Musk Attacks Free-Speech" scandal on the site formerly known as Twitter.

No, I'm not talking about the now-X'd Liam Nissan parody account.

The Liam Nissan affair—the perception that a parody account was banned for humiliating Dear Leader (aka "Space Karen")—was bad enough. The reality, according to a Bluesky version of "Liam Nissan," is that he (or she) chose to quit X because of all the threats (s)he was getting.

Today, the news on X was that the world's most famous exploding carmaker had suspended the account of @yulia_navalnaya, Alexei Navalny's widow.

The scandal gave the account in question instant global publicity. The account has now been restored and is amassing followers by the tens of thousands.

Is it really her, though?

The recently launched Yulia Navalnaya account is all in Russian. It looks real. There's even a video.

As a social media user/news consumer, I can only assume that journalists and others close to Navalnaya have vouched for the authenticity of this account and its content.

But as I watched it with my own eyes, I realized I had no way of knowing if it was real or not.

Why? Because in this age of deepfakes, Russian misinformation campaigns, untrustworthy $8-a-month "verification," combined with Elon Musk's fascist-friendly nature and frequently repeated desire to overthrow democracy, it's impossible to be 100% sure about anything.

Celebrities are already sliding into the DMs of AI-generated influencers like @fit_aitana.

And the way Elon Musk is running X isn't helping.

AI videos are now too real to trust

The picture at the top of this article is a screenshot from a video created with Sora, OpenAI's new video generation model.

The scene was created with a simple text prompt: "A litter of golden retriever puppies playing in the snow. their heads pop out of the snow."

As the Feb 19 edition of The AI Breakdown podcast details, there has been a strong negative reaction to this latest AI "advance," with comments including:

"Gen AI is a fucking insult to humanity. Image and video are no longer historical documents. With AI, images will become nothing more than our entire visual history shoved into a meat grinder and served to us meaninglessly for profit. A craft I’ve dedicated my entire life to is being replaced by an image predictor slot machine"—@meganroseruiz

"I’m struggling to think of a single positive thing making realistic ai generated videos like this will bring. It’s all just net negative and dystopian"—@MoistCr1TiKaL

On TikTok, users including @allyrooker are speaking out against the many ways people will be "using this for evil," from deepfake porn and sexual exploitation to political corruption. @megan_b_rice says this goes beyond art and human creativity. Once "governments and military" start using it, "we would not begin to even know what is the truth."

We're in a new age of hyper-realistic deepfakes that, combined with voice cloning, will make it easy for anyone with enough "compute" power to create deceptive videos that could go viral in minutes.

What uses might evildoers and authoritarians find for hyper-realistic AI video?

OpenAI's Sora, which generates up to 60 seconds of video from a single text prompt—with the possibility to extend the same video one minute at a time—makes it easier than ever to imagine AI video being used for ongoing propaganda purposes and election interference.

But it could also be used to manipulate financial markets, create fake criminal accusations and courtroom evidence, mislead people about basic science and facts, smear public figures or corporate executives, create panic by making terrorist threats seem more real, and a whole lot more.

As more of the public absorb the downsides of AI—including the massive impact on jobs—the backlash will likely grow, especially as people realize that, as The AI Breakdown states, a "tiny handful of companies" are "making decisions about these things that will have such a dramatic impact on their lives."

Just as journalism accelerates toward an "extinction-level event," AI video is creating an "Oppenheimer moment" in the war on truth.

It's not too far-fetched to think a new account could suddenly appear on social media purporting to be a famous figure and for that account to deliver a realistic video-and-voice message that tens of millions could see, hear—and act on—without ever knowing if it's real or fake.

While paid subscriptions and one-time tips are always welcome, all content to this ad-free newsletter is free, so subscribe for $0 and you won’t miss anything. And feel free to share this content knowing there’s never a paywall!