Updated on April 11, 2026 at 10:15
© Berke Citak (Unsplash)
This Frenchman wanted to « test AI surveillance », but he certainly didn't expect that. The statements directed at an AI quickly mobilized the FBI and crossed the Atlantic. The case ended with the RAID intervention at the home of the thirty-year-old man in Strasbourg.
On Friday, April 3, 2026, a 37-year-old Strasbourg resident was apprehended by the RAID, the elite unit of the National Police, according to Les Dernières Nouvelles d’Alsace. It was a simple conversation with an AI that led the authorities to the thirty-year-old man.
The name of the artificial intelligence in question has not been disclosed.
When questioned by the authorities, the man explained that he had wanted to test the reliability and monitoring of the artificial intelligence. In this context, he told his chatbot that he had planned to buy a firearm, a Glock, in order to kill a CIA, Mossad, or DGSI intelligence agent. Even though the threat was clearly fictitious, it caught the attention of the FBI. The American federal bureau was likely informed by the developer of the AI.
American intelligence services turned to Pharos, the official French platform for reporting illegal content or behavior online to the police and gendarmerie. Based on the information provided by the FBI, law enforcement decided to intervene immediately. That is why the RAID stormed the man's home on Friday afternoon. The man was released from custody the following day, and the case was closed without further action. No weapon was found at his home during the RAID's visit. Due to his psychiatric history, the detained individual was hospitalized.
Also see : ChatGPT and the military – after a controversial agreement with the Pentagon, an OpenAI leader resigns by « principle »
Data shared in case of danger
OpenAI, the AI giant developing ChatGPT, does not shy away from « sharing your personal data, including information related to your interaction with our services, with public authorities », in order to « protect the security, integrity, and safety of our products, employees, users, or the public ». More broadly, any violation of the terms of use can lead to data sharing with the relevant authorities
.Same tune on the side of other AI giants. Anthropic, the startup behind Claude, explains that it does not disclose « information about its customers or end-users of its services in response to government requests, except in accordance with a valid legal procedure », or if « there is an emergency that could lead to an imminent danger of death or serious injury, and the disclosure of such information could prevent this danger ». Unlike OpenAI, Anthropic nevertheless commits to challenging requests deemed too broad or vague. Same tune on the Google side, which can share messages exchanged with Gemini in order « to protect user and public safety ».
👉🏻 Follow real-time tech news: add 01net to your Google sources, and subscribe to our WhatsApp channel.
Source: Les Dernières Nouvelles d’Alsace
- Google Play Store
- 4K Video Downloader Plus
- VLC Media Player
- Adobe Reader
- uTorrent
- WinRAR
- Xender
- Google Chrome
- CCleaner
- Mozilla Firefox
- Avast Free
- ChatGPT
- Microsoft 365
- Opera One
- Avast One
Download
- Google Play Store
- 4K Video Downloader Plus
- VLC Media Player
- Adobe Reader
- uTorrent
- WinRAR
- Xender
- Google Chrome
- CCleaner
- Mozilla Firefox
- Avast Free
- ChatGPT
- Microsoft 365
- Opera One
- Avast One
