From bunnies jumping on a trampoline at night to Norwegian soccer player Erling Haaland getting scared by his own reflection, the more that AI-generated content makes its way into our social feeds, the harder it becomes to differentiate what’s false from what’s real.
And while the rise of skepticism is unsurprising, it’s proving to be an even bigger problem when the stakes are high.
Following the death of Republican Sen. Lindsey Graham of South Carolina on Saturday, users flocked to social media to share memes and speculation about the health of another Republican—Mitch McConnell. The senator from Kentucky, who was admitted to a hospital due to a fall, had not been seen since Graham’s death announcement, fueling rumors that he had also passed away.
His team proceeded to try and prove the senator’s well-being the traditional way: issuing a statement and releasing a recent photo of McConnell in seemingly good health.
But for an online public that’s grown accustomed to being tricked by AI, photographic evidence was deemed insufficient. Instead, it’s sparked its own rabbit hole of responses.
What’s a picture even worth?
Some of the initial reactions poked fun at the situation. Take late-night talk show host Jimmy Kimmel, who posted a parody photo substituting McConnell’s face for his own. The caption on Kimmel’s official Instagram read: “For those who’ve been asking, I’m feeling great.”
But for others, the tone of the reaction was more conspiratorial, with many users online believing the photo to be fake or AI-generated.
“Mitch McConnell may or may not be alive. But his staff are definitely dying a slow death watching the AI photo they released yesterday get ripped to shreds,” a user said on X, despite no credible evidence of the image being AI.
Many users paid special attention to the newspaper next to McConnell, with fake images circulating that featured warped text and gibberish, as is common with AI-generated images.
However, the original image posted by McConnell’s team does not actually look like that. Additionally, the newspaper has since been identified as the Sunday sports section of The Washington Post.
Turning to AI for a ruling
Perhaps ironically, other skeptics turned to AI itself to try and verify the image.
“@grok, analyze this image for Ai, also focus on the image in the newspaper and find it, if real, when it was published and from what newspaper?” one user said on X, tagging the in-app AI chatbot.
Grok responded: “The photo is AI-generated (SynthID watermark detected by OpenAI tool; confirmed fake by Snopes, Cincinnati Enquirer, Courier Journal). McConnell’s office released no such image—pure hoax.”
The AI bot continued: “The newspaper (and photo inside) is also AI-fabricated. No real paper or date matches; entire scene is synthetic,” despite the newspaper spread actually existing.
When challenged on the conclusion by the original user, who provided the newspaper spread to Grok, the AI stood behind its original—albeit false—judgment.
Others turned to Google Lens for clarity, uploading McConnell’s photo to the search engine, with the AI summary returning a hallucinated response, indicating the image was taken in 2023.
Skepticism around McConnell’s health state and the response to it are not entirely surprising. Public trust in the government is among the lowest it’s been in seven decades, with only around 17% believing that Washington will “do what is right,” according to Pew Research.
To an already distrustful public, the way McConnell’s team has navigated the senator’s hospital stay—in secrecy—lends itself as a catalyst for theories, a similar treatment that then-President Joe Biden received when he was seeking reelection amid growing age and health concerns in 2024.
But as public trust continues to erode, the widespread use of convincing AI images adds a new layer of complexity.
While unfounded conspiracy theories have long been commonplace online, (remember Pizzagate?), the recent and rapid discourse arising from McConnell’s photo is symptomatic of a more recent trend—in which users are fatigued and skeptical of AI content, yet reliant on it at the same time.
“Funny memes about Mitch McConnell aside, the replies are a mess,” one user noted on X. “The fact that people have to ask Grok if this photo is AI, and then they base their entire worldview on what the AI says, is not a good sign of what’s to come.”
