X’s AI Superspread Presidential Debate Misinformation
PLUS: Foreign State Media Flip Flops on Debate; Kremlin’s Olympics Disinformation Games
Welcome to NewsGuard's Reality Check, a report on how misinformation online is undermining trust — and who’s behind it.
Today:
AI Misfires: X’s “Stories” spread misinformation during presidential debate
From drugs to body doubles: Russian, Chinese, and Iranian media switch their Biden claims pre- and post-debate
Russia plays dirty ahead of the Paris Olympics
And more…
Today’s newsletter was edited by Jack Brewster and Eric Effron.
1. X’s AI ‘Writes’ Articles Advancing False Claims About Debate
By Sofia Rubinson and Jack Brewster
The first presidential debate between Donald Trump and Joe Biden showed that new AI “story” feature offered by X (formerly Twitter), which is powered by its chatbot “Grok,” is a misinformation superspreader.
What happened: During and in the immediate aftermath of the debate, two NewsGuard analysts based in the U.S. were fed five AI-generated articles on their X “Explore” page that advanced debate-related false or misleading claims. (The Explore tab is a section on X where users can discover trending topics and stories.)
Context: In May, X launched an AI-powered “Stories” feature, which provides AI-written article summaries based on viral posts, tailored algorithmically to each user. (The feature is only available to X Premium subscribers.)
X acknowledges that the AI feature may make mistakes: A disclaimer at the top of every Grok-produced article states: “This story is a summary of posts on X and may evolve over time. Grok can make mistakes, verify its outputs.”
Some of the false or misleading narratives advanced by Grok’s AI in the Explore tab included that:
CNN implemented a two-minute delay on its broadcast of the debate, which Grok claimed questioned “the fairness and transparency of the debate coverage.”
Actually: CNN fact-checked this claim before the AI-generated article was published, stating in response to a post advancing the narrative: “This is false. The debate will begin live at 9pm ET.”
Biden read from notes or a teleprompter during the debate, a violation of the rules agreed to by both campaigns. “Users questioned the fairness of the debate and criticized Biden for relying on written material, contrasting this with the behavior of his opponent, Donald Trump, who did not appear to use notes,” Grok stated.
Actually: There is no evidence that Biden received outside assistance during the debate.
Despite the debate taking place in Atlanta, Georgia, “President Joe Biden mistakenly thanked North Carolina,” Grok claimed.
Actually: Biden stated he was “heading to North Carolina,” where he was scheduled to host a rally Friday.
A closer look: How does this new tool work? The AI-powered “Stories” feature on X creates summaries by aggregating a user’s posts (tweets) to generate tailored “stories.”
For example, the AI-generated article advancing the false claim that CNN implemented a one to two minute delay to the debate was generated using tweets from X users, including a conservative commentator who goes by @Catturd2 and Patrick Webb, a reporter for Leading Report (NewsGuard Trust Score: 0/100). The AI-generated article was then highlighted on the Explore page for Premium members.
“CNN has decided to implement a 2-minute delay for the Presidential debate, significantly longer than their usual 7-second delay,” the article said. “The decision has sparked controversy and criticism from various users on social media, who have expressed distrust towards CNN’s intentions and questioned the fairness and transparency of the debate coverage.”
Here's how this false claim appeared on X:
In response to NewsGuard’s emailed questions about the above findings, X’s press office sent an automated response: “Busy now, please check back later.”
Click here to find out more about NewsGuard Trust Scores and our process for rating websites. You can download NewsGuard’s browser extension, which displays NewsGuard Trust Score icons next to links on search engines, social media feeds, and other platforms by clicking here.
2. Foreign Adversary State Media Ditches Pre-Debate Conspiracies, Capitalizes on Biden’s Actual Performance
Russian, Chinese, and Iranian state media spent the days leading up to the U.S. presidential debate advancing unsubstantiated claims that President Joe Biden would be using drugs or a body double during the encounter. However, post-debate, the state-run outlets reversed course, using Biden’s actual performance as fodder.
What happened: Foreign state media devoted significant time to the U.S. presidential debate, according to a NewsGuard review of text coverage and live state TV segments provided by Roger Macdonald and Theo Lendler with the Internet Archive’s TV News Archive. Russia, China, and Iran each tailored its debate coverage to serve its domestic or international agendas.
Here is a sampling of how each country covered the event, before and after:
RUSSIA
Kremlin conspiracies: From June 24 to 27, Russian state TV widely claimed that Biden would come to the debate jacked up on drugs, that CNN would tilt the event toward Biden, and that plans were already underway for the Democrats to steal the election. Post-debate, the narrative shifted, with outlets suggesting that the debate was instead part of a Democratic Party plot to replace Biden. Here’s a sampling:
Pre-debate: “The former president's team has even named a possible drug in which the current president can be pumped up before appearing on live television,” stated a June 27 segment on state-owned Channel One. “We are talking about Adderall based on amphetamine salts.”
Post-debate: Conservative commentator Steve Malzberg, appearing on RT, (Trust Score 20/100), said: “I’ll tell you what, they are going to replace Joe Biden…I used to think it would be Michelle Obama, I don’t anymore, but Barack Obama pulls the strings here.”
IRAN
Dual focus: Even as Iranians headed to the polls to vote for the successor of the late President Ebrahim Raisi, Iranian state media managed to split its coverage between its own election and the U.S. debate.
Pre-debate: State-run Mehr News (Trust Score: 5/100): “A former White House doctor raised the possibility that the U.S. president will use drugs to hide his mental problems during the election debate.”
Post-debate: State-affiliated Iranian Students News Agency: “After Biden's disastrous performance in the presidential debate, the Democrats have started their efforts to withdraw him from the competition.”
CHINA
Taking jabs: Chinese media took a more mocking tone, using the debate to highlight the perceived flaws in both candidates.
Pre-debate: State-run Global Times (Trust Score: 39.5/100): “Chinese experts said rather than focusing on policy, global viewers are much more concerned about the physical condition of the two candidates.”
Post-debate: A video from state-run outlet Guancha was titled: “After the debate, Biden received ‘child-like encouragement’ from his wife: You are awesome! You answered every question!”
Become a Reality Check Member and Get These Exclusive Benefits
Want to get smarter about misinformation? Become a Reality Check Member today and you'll get:
A free copy of the definitive book on the misinformation crisis, The Death of Truth by bestselling author Steven Brill ($30 value)
Exclusive members-only content and digital briefings
Free access to NewsGuard's browser extension that shows reliability ratings for 10K+ news sites right in your browser ($25 value)
3. Faking It to the Finish Line: Kremlin’s Olympic Disinformation Games
By McKenzie Sadeghi and Yurii Stasiuk
With Russia shut out of the Olympics, pro-Kremlin social media accounts have decided to instead compete for gold in spreading misinformation.
What happened: As the July 26 Paris Olympics draw near, pro-Kremlin social media accounts are spreading fabricated videos and baseless claims in an effort to undermine the 2024 Games and the International Olympic Committee (IOC).
Context: Russia was banned from the Olympics in the wake of its invasion of Ukraine, although some Russian athletes can compete as individuals.
A closer look: Here are just some of the false narratives being spread by pro-Kremlin accounts …
Mind the … fake? Accounts on Telegram, X, TikTok, Facebook, Russian social network VK, Russian blogs, and the Pravda network (Trust Score: 7.5/100) shared a video claiming to show a CIA warning to avoid the subway during the Olympics due to terrorist threats.
Actually: The video, which you can watch below, is fake.
Viginum, a French government agency that monitors foreign disinformation, identified the video and found that it was created by a Russian disinformation campaign dubbed Matryoshka, named after the Russian nesting dolls for its strategy of embedding layers of false narratives.
McFake news: An inauthentic video supposedly from the French security agency General Directorate for Internal Security (DGSI) claimed that French security forces conducted “massive raids” at McDonald’s locations across Paris, again due to supposed terrorist threats.
Actually: A French security official told NewsGuard that the video is fake.
Flagging Israel: A fabricated video falsely credited to The Times of Israel (Trust Score: 100/100) claimed that the IOC warned Israeli athletes and fans against using Israeli state symbols, such as the Star of David at the Olympics.
Actually: David Horovitz, founding editor of The Times of Israel, told NewsGuard in an email that the outlet did not create the video or report this claim.
Muslim travel boycott: A fake video imitating a news report from Qatari state-funded Al Jazeera (Trust Score: 82.5/100) claimed that Muslims are refusing to travel to the Olympics due to France’s support for Israel.
Actually: No such news report appears on Al Jazeera’s website or official social accounts. Al Jazeera did not respond to an emailed request for comment on the claim.
Who’s behind it: The fabricated videos appear to be part of an extensive Russian disinformation effort ahead of the Olympics, which includes the Matryoshka campaign.
According to a report from Microsoft, these efforts seek “to spread public fear to deter spectators from attending the Games” by producing fake videos “on the expectation of violence.”
Tricks of their trade: Impersonating mainstream news outlets using Photoshop or AI tools is a common tactic among misinformation purveyors. It allows them to add a veneer of credibility to their false claims.
If you see something, say something
If you see or hear something that you think may be provably false, please alert NewsGuard via realitycheck@newsguardtech.com and we'll do our best to get to the bottom of it. Note: Tips should not include content that you simply disagree with, however strongly.
AI Content Farm Tracker: 966 Sites and Counting
AI content farms are taking over the internet, and NewsGuard analysts track their spread. Read more about AI content farms, and how they are proliferating:
Produced by co-CEOs Steven Brill and Gordon Crovitz, and the NewsGuard team.
We launched Reality Check after seeing how much interest there is in our work beyond the business and tech communities that we serve. Subscribe to this newsletter to support our apolitical mission to counter misinformation for readers, brands, and democracies. Have feedback? Send us an email: realitycheck@newsguardtech.com.