Elon's Next Attack

We can't trust anything that comes out of Donald Trump's mouth. Or out of Elon Musk's Grok AI.

How it started: In May 2024, Elon Musk told us that Grok AI was going to be the future of news.

How it's going: In August 2024, Grok AI users on X are now sharing realistic-looking images of Elon Musk as a school shooter, leaving huge trails of blood at Sandy Hook Elementary School, Uvalde Elementary School and other places. Sometimes Elon is acting alone. Others, he has accomplices such as Mickey Mouse and the video game character Mario.

"This is insane."

X user and Substacker Oliver Alexander writes: "This is insane. You don't even need to confuse the AI to generate these images. Releasing this as an actual paid product, without any safety net, is going to result in an expensive legal battle for X and Elon Musk."

Of course, fun-loving X "premium" subscribers (who were given beta access to the new Grok AI image generator this week) quickly began testing their boundaries of Musk's "anti-woke" (and "completely unregulated") new picture-making tool.

This image of Trump (who told Elon this week that based on this week's TIME cover, "Camilla" looked like "the most beautiful actress ever to live... very much like our great first lady, Melania") kissing Kamala Harris quickly began making the rounds.

Trump can also be seen not just kissing Elon Musk but also celebrating a pregnancy with him:

As Oliver Alexander also wrote on X: "It is one thing to have completely unregulated AI image models that can run locally. It is something entirely different to tie it directly to your multibillion dollar social media platform. Who on earth allowed this to be pushed out to the public like this?"

In May, I wrote that Grok News had been "launched amid war in an election season rife with mis- and disinformation" and had "gone live without waiting to figure out 'issues like citation and hallucination'."

Reminder: This newsletter is free to all thanks to the support of my paid subscribers... if you'd like to upgrade, here's a special offer link:

Elon's Grok "News" product is designed to effectively steal the real-time reporting of actual journalists and merge it with the slanted viewpoints of his paid users (an audience which skews heavily toward radicalized Musk fanboys).

Now, less than three months before Election Day, Musk is giving these fanboys a potentially even more powerful weapon with which to spread propaganda and sow confusion.

Visuals deceive even more than words

As experts have long noted, our brains are wired to trust visuals (as well as people with large Twitter followings, like Musk himself). Inserting doctored or fake images into news-related posts greatly increases the chance they will be believed—allowing them to create a lasting impression that may be hard to shake.

Just two weeks ago, Musk himself was slammed for sharing a parody "deepfake" video of Kamala Harris.

This week, Musk welcomed Donald Trump, the world's most notorious lie-spewer, back to X. Since then, they've both shared an AI-generated dance video starring themselves on their separate feeds, which now have a combined total following of more than 284 million, which (if you don't correct for bots or duplication) makes them bigger than Indonesia, the 4th largest country by population in the world.

Given Trump's recent fixation with AI, it's hard to imagine—as both he and Musk become more desperate about a race that's slipping away from Republicans—that they will limit themselves to sharing only the most fun user-generated content they see.

Musk has weaponized Grok AI in his "battle to the death" with America

As is now obvious do anyone paying attention, when Elon declared in March 2024 "this is a battle to the death with the anti-civilizational woke mind virus," he wasn't kidding.

Five months later, Elon continues to casually tweet lines like "civil war is inevitable" as he stokes race riots in the UK, attacks traditional media in the US, and joins Trump in spreading doubt about the fairness of the upcoming election.

A recent analysis showed that Musk's false claims about the upcoming US election had already generated more than 1.2 billion impressions.

Meanwhile, Musk has also:

• Funded a Pro-Trump PAC that is deceptively stealing swing-state voter data
• Hosted a two-hour "conversation" with Trump on X that was also an illegal campaign donation
• Broken US labor laws when part of that conversation involved actively intimidating American workers
• Been sued by a group of Secretaries of State in key swing states after Grok was found to be consistently spreading disinformation about Harris missing the ballot deadline in multiple states

This week, Minnesota Secretary of State Steve Simon talked to Ali Velshi on MSNBC about the lawsuit, the easy fix that ChatGPT has already applied, but which Musk has resisted—and the fact that at least two Secretaries of State have been "swatted" in recent days as a result of the efforts of people like Trump and Musk to use mis- and disinformation to activate MAGA cultists who prefer violence to reason.

Musk's Saudi-backed takeover of Twitter has been disastrous from a business sense—and has, as expected, forever ruined the reputation of the Musk-hired X CEO Linda Yaccarino.

But the bloodthirsty Saudis aren't complaining. Clearly satisfied by the unpublicized side benefits of their money-losing investment in X, the free-speech-squashing Saudis also shoveled money into the recent (and massive) $6 billion funding round which valued xAI at $24 billion.

Despite the conventional wisdom being spouted by the likes of the New York Times that AI hasn't been as prominent a factor in the election as people expected, it's important to look further than the organ that brought you the classic October 2016 story: "Investigating Donald Trump, F.B.I. Sees No Clear Link to Russia."

As Parmy Olson explained recently in Bloomberg Opinion: The most viral forms of disinformation are in fact the easiest to spot and counter. It's the ability to say "the same thing, slightly differently, thousands of times... (that) makes AI-powered disinformation so much harder to detect."

We don't yet know how many elections AI will tilt in one direction or another in 2024.

But we do know some things for certain:

We can't trust anything that comes out of Donald Trump's mouth.

And we can't trust anything that comes out of Elon Musk's Grok AI.

Thanks for reading! While paid subscriptions (billed to TLD Media LLC) 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.