Misinformation

When researchers and journalists searched Google for Philip Morris International, British American Tobacco, and James Hardie — companies with decades of documented harm — Google's AI Overviews returned glowing summaries drawn heavily from the companies' own websites. Philip Morris was described as "a leading international tobacco company working to transition from cigarettes to smoke-free products" focused on "a future without cigarettes," with no mention of court findings that the company spent decades lying to the public about the addictiveness and health risks of smoking. James Hardie, once Australia's largest asbestos distributor, was hailed as a "global leader" that "pioneered asbestos-free fibre cement" — omitting that asbestos products still kill thousands of Australians each year.

University of Sydney public health professor Becky Freeman called the Philip Morris summary "essentially a regurgitation of Philip Morris International's PR materials." Experts say companies are now racing to optimise their websites specifically so AI systems ingest and repeat their preferred narratives — a practice known as generative engine optimisation (GEO). Google maintains it does not allow paid influence over AI Overviews and draws from sources it deems most reliable, but its own disclaimer acknowledges the feature "may sometimes provide inaccurate content." After the ABC contacted Google, subsequent searches began returning overviews that at least noted Philip Morris's "Big Tobacco" classification — though Google denies that change had anything to do with the inquiry.

A new study found that AI chatbots — including ChatGPT and its rivals — served up incorrect election information nine times out of ten. For tools that millions of people increasingly turn to for quick answers, that's a remarkable batting average in the wrong direction.

The findings land at a particularly awkward moment: AI companies have been eager to position their products as trusted information assistants, while researchers keep finding that when it comes to high-stakes civic topics, these tools are less "knowledgeable friend" and more "confidently wrong acquaintance."

Almira Osmanovic Thunström, a researcher at the University of Gothenburg, fabricated a skin condition called bixonimania and seeded it across the internet via a fake university, a fake researcher named "the Lying Loser" (in Croatian), and a preprint paper funded by "the Galactic Triad" and thanking Professor Ross Geller. She expected human moderators or AI filters to catch it. They did not. Multiple popular AI chatbots began suggesting bixonimania as a possible diagnosis for users describing eye discomfort after screen use.

Worse, the fake paper was cited in a real peer-reviewed journal, which only boosted the condition's apparent legitimacy in AI training data. The experiment illustrates how thin the line is between "information on the internet" and "medical fact" as far as large language models are concerned — and how little it takes to cross it maliciously.

Steven Rosenbaum's The Future of Truth: How AI Reshapes Reality — a buzzy book about how AI distorts reality — turns out to contain more than a half-dozen misattributed or fabricated quotes, discovered by a New York Times review. Rosenbaum admitted he used ChatGPT and Claude during research and writing, and is now working with editors to correct the affected passages.

Among the hallucinated quotes was one falsely attributed to tech journalist Kara Swisher, placed in a chapter about AI lies, in which she supposedly mused that language models are "like a mirror" that reflects "our own morality back at us, polished and articulate, but ultimately empty behind the surface." The real Swisher told the NYT she "never said that," adding that the quote made her "sound like I have a stick up my butt, according to ChatGPT." Rosenbaum, who holds a self-described "Masters Degree in Truth" from NYU, suggested the whole fiasco was a teaching moment — which is one way to describe writing the book on post-AI truth while not fact-checking your AI-generated quotes.

An anonymous conceptual artist posted a cropped image of an authentic 1915 Claude Monet Water Lilies painting — currently hanging in Munich's Neue Pinakothek — with a caption claiming it was AI-generated, and helpfully slapped X's official "Made with AI" label on it for good measure. The internet delivered: 6.7 million views' worth of confident critique followed, with commenters calling it a "cluttered slop" that achieves "like 20% of it" and dissecting its supposedly incoherent colors and missing depth. One person filed a 700-word autopsy of the fake. The painting was not fake.

The mass delete button got a workout once the reveal landed, but screenshots are forever. The episode neatly illustrates what researchers already knew: people's perception of art quality shifts dramatically based on told source rather than actual quality. As one observer put it, most people's judgment about whether something is AI is "wrong and biased by its source." Turns out the real threat to art criticism wasn't AI — it was a label.

A 2025 EY Canada report on loyalty fraud — Points of Attack: Uncovering Cyber Threats and Fraud in Loyalty Systems — turns out to be a masterclass in what GPTZero calls "vibe citing": letting an LLM hallucinate your references so you don't have to. An investigation found broken URLs, nonexistent Gartner documents, a fabricated McKinsey report, and a Forbes article that doesn't exist — all in a 44-page publication credited to two partners and a senior manager at one of the world's biggest consulting firms.

The damage doesn't stop at embarrassment. The report has since been cited in a Canberra Times article syndicated to over 60 Australian newspapers, and Claude, ChatGPT, and Perplexity have all been caught surfacing its hallucinated claims in response to queries. A fake McKinsey citation was apparently laundered from an obscure fintech blog, which itself made up the source — a game of telephone where every player is an AI and nobody checks the tape.

A video circulating on Threads claimed to show a U.S. Navy warship ablaze following an Iranian attack — complete with a caption insisting "This is not AI." It was, in fact, very much AI. The Hive Moderation detection tool rated it 84.4% likely AI-generated, and the telltale signs were all there: unnaturally smooth fire, water that refused to react to a supposed inferno, and ships doing the subtle warp-and-flicker dance that AI video generators still haven't figured out.

For context, while Iran did target USS Truxtun, USS Rafael Peralta, and USS Mason in early May 2026, U.S. Central Command confirmed none of the vessels were hit. No major news outlet reported any warship being damaged — a rather significant detail that AI-generated doom-scrolling content is under no obligation to acknowledge.

A syndicated 64-page summer supplement published in the Chicago Sun-Times and Philadelphia Inquirer included a summer reading list full of books that simply do not exist — titles and summaries alike conjured entirely by AI. Freelance writer Marco Buscaglia, working for King Features Syndicate (a Hearst unit), admitted he "just kind of republished this list that [an AI program] spit out" without fact-checking it, exposing two major newspapers to widespread public mockery.

Buscaglia accepted blame and his subsequent firing, while King Features claimed his AI use violated their "strict policy." The Sun-Times' parent, Chicago Public Media, pulled the section from its e-paper, promised refunds to print subscribers, and launched an investigation into other content in the section — since Buscaglia acknowledged using AI elsewhere in the supplement and couldn't guarantee any of it was properly vetted. The Sun-Times Guild summed it up neatly, condemning the "slop syndication."

The Mississippi Free Press published an opinion column on April 7 before discovering it was AI-generated and its supposed author was entirely fictional — fake name, fake headshot, fake social media profiles, and a résumé linking to a company whose editor had already caught the same person running the same con. The ruse only unraveled when the phantom columnist submitted an invoice that didn't match their name.

Further digging revealed a "raft" of similar AI-generated submissions from other fake authors, apparently originating outside the country — none of which were published. The outlet has since pulled three additional queued columns showing similar red flags, and is now working on a formal AI policy, staff training, and tighter vetting. As for AI detectors: the editor notes they're not very reliable, which is a fun problem to have when the fakes are getting better every day.

Two days before the New Hampshire presidential primary, voters received a robocall featuring a convincing AI-generated imitation of President Biden's voice — urging them to skip the primary and "save your vote for November." The call even opened with Biden's signature "What a bunch of malarkey!" before proceeding to deliver exactly that. The Biden campaign confirmed he recorded no such message, and the New Hampshire attorney general launched an investigation, calling it an apparent felony-level attempt to suppress votes.

The call's callback number — a clever touch — was routed to Kathleen Sullivan, treasurer of a PAC encouraging Democrats to write Biden in on the ballot. Sullivan, suddenly fielding mysterious calls, connected the dots and filed a complaint with the AG's office. In New Hampshire, using fraudulent information to deter voting is a felony. The attorney general helpfully reminded voters that participating in the primary would not, in fact, prevent them from voting in November — a clarification that shouldn't need to be made, and yet, here we are.