A City Council Candidate Made Fake CNN News Stories on His Phone. It Probably Won't Be the Last Time.
A New York race showed how easy it is to generate convincing fake political news with AI. Experts say the tools are only getting cheaper and faster.
Key points
- Jonathan Rinaldi, a Queens city council candidate, used an AI chatbot in October 2024 to generate fabricated news stories and fake endorsements about his opponent.
- One fake story carried a CNN logo and falsely claimed incumbent Lynn Schulman had dropped out of the race.
- Rinaldi shared the fabricated content on Facebook and Instagram during the active campaign period.
- Schulman won the November 2024 election by a landslide despite the false claims.
- Researchers and election-integrity experts told The Guardian that AI-generated political disinformation is spreading faster than platforms can respond.
Jonathan Rinaldi was lying in bed, tapping on his iPhone, asking an AI chatbot to invent news stories for him. He was running for a city council seat in Queens, New York. His opponent was Lynn Schulman, a sitting Democrat. The stories he generated were fake. He shared one of them anyway.
The post looked real. It carried a CNN logo. It claimed Schulman had been "forced to drop out of the race due to a series of critical mistakes." She had not dropped out. She had not made any such mistakes. She won the election in November 2024 by a landslide.
How hard is it to make a fake news story like this?
Not very hard at all. Consumer AI chatbots can draft convincing text in seconds. Adding a recognisable logo takes basic image editing, or sometimes nothing more than another AI prompt. The whole operation, as Rinaldi demonstrated, fits on a phone screen.
This is the part that worries researchers. The barrier used to be skill: you needed to know how to edit images, write convincingly, and build distribution. Now the barrier is closer to zero.
Deepfakes, meaning AI-generated photos or videos manipulated to show real people doing or saying things they never did, add another layer. A fabricated clip of a candidate conceding, or making a controversial statement, can travel across social media in hours. By the time a denial goes out, millions of people may have already seen the original lie.
For ordinary voters, the practical upshot is uncomfortable but simple. If a dramatic political story appears in your feed, especially one involving a candidate quitting or being caught in a scandal, slow down before you share it. Search the candidate's name on a few different sources. Check whether any established news organisation is reporting the same thing. A story that exists only as a screenshot, or only on one obscure page, is a warning sign.
Rinaldi's stunt failed. Schulman won easily. But the next attempt may be better timed, better made, or aimed at a race where the margin is thinner. The tools will not get harder to use.
