Paris prosecutors raid X offices as part of investigation into child abuse images and deepfakes

PARIS (AP) — French prosecutors raided the offices of societal media level X connected Tuesday arsenic portion of a preliminary probe into allegations including spreading kid intersexual maltreatment images and deepfakes. They person besides summoned billionaire proprietor Elon Musk for questioning.

X and Musk's artificial quality institution xAI besides look intensifying scrutiny from Britain's information privateness regulator, which opened ceremonial investigations into however they handled idiosyncratic information erstwhile they developed and deployed Musk's artificial quality chatbot Grok.

Grok, which was built by xAI and is disposable done X, sparked planetary outrage past period aft it pumped retired a torrent of sexualized nonconsensual deepfake images successful effect to requests from X users.

The French probe was opened successful January past twelvemonth by the prosecutors’ cybercrime unit, the Paris prosecutors' bureau said successful a statement. It's looking into alleged “complicity” successful possessing and spreading pornographic images of minors, sexually explicit deepfakes, denial of crimes against humanity and manipulation of an automated information processing strategy arsenic portion of an organized group, among different charges.

Prosecutors asked Musk and erstwhile CEO Linda Yaccarino to be “voluntary interviews” connected April 20. Employees of X person besides been summoned that aforesaid week to beryllium heard arsenic witnesses, the connection said. Yaccarino was CEO from May 2023 until July 2025.

A spokesperson for X did not respond to aggregate requests for comment. X's lawyer successful France, Kami Haeri, told The Associated Press: ″We are not making immoderate remark astatine this stage."

In a connection posted connected X, the Paris prosecutors’ bureau announced the ongoing searches astatine the company's offices successful France and said it was leaving the level portion calling connected followers to articulation it connected different societal media.

“At this stage, the behaviour of the probe is based connected a constructive approach, with the purpose of yet ensuring that the X level complies with French law, arsenic it operates connected the nationalist territory,” the prosecutors' connection said.

European Union constabulary bureau Europol "is supporting the French authorities successful this," Europol spokesperson Jan Op Gen Oorth told the AP, without elaborating.

French authorities opened their probe aft reports from a French lawmaker alleging that biased algorithms connected X apt distorted the functioning of an automated information processing system.

It expanded aft Grok generated posts that allegedly denied the Holocaust, a transgression successful France, and dispersed sexually explicit deepfakes, the connection said.

Grok wrote successful a wide shared station successful French that state chambers astatine the Auschwitz-Birkenau decease campy were designed for “disinfection with Zyklon B against typhus” alternatively than for wide execution — connection agelong associated with Holocaust denial.

In aboriginal posts connected X, the chatbot reversed itself and acknowledged that its earlier reply was wrong, saying it had been deleted and pointed to humanities grounds that Zyklon B was utilized to termination much than 1 cardinal radical successful Auschwitz state chambers.

The chatbot besides appeared to praise Adolf Hitler past year, successful comments that X took down aft complaints.

In Britain, the Information Commissioner’s Office said it’s looking into whether X and xAI followed the instrumentality erstwhile processing idiosyncratic information and whether Grok had immoderate measures successful spot to forestall its usage to make “harmful manipulated images.”

"The reports astir Grok rise profoundly troubling questions astir however people’s idiosyncratic information has been utilized to make intimate oregon sexualised images without their cognition oregon consent, and whether the indispensable safeguards were enactment successful spot to forestall this,” said William Malcolm, an enforcement manager astatine the watchdog.

He didn’t specify what the punishment would beryllium if the probe recovered the companies didn't comply with information extortion laws.

A abstracted probe into Grok launched past period by the U.K. media regulator, Ofcom, is ongoing.

Ofcom said Tuesday it's inactive gathering grounds and warned the probe could instrumentality months.

X has besides been nether unit from the EU. The 27-nation bloc's enforcement limb opened an probe past period aft Grok spewed nonconsensual sexualized deepfake images connected the platform.

Brussels has already deed X with a 120-million euro (then-$140 million) good for shortcomings nether the bloc's sweeping integer regulations, including bluish checkmarks that broke the rules connected “deceptive plan practices” that risked exposing users to scams and manipulation.

On Monday, Musk ’s abstraction exploration and rocket business, SpaceX, announced that it acquired xAI successful a woody that volition besides harvester Grok, X and his outer connection institution Starlink.

___

Associated Press writers Nicolas Vaux-Montagny successful Lyon, France, Mike Corder successful The Hague, Netherlands, Sylvia Hui and Kelvin Chan successful London contributed to this report.

Copyright 2026 The Associated Press. All rights reserved. This worldly whitethorn not beryllium published, broadcast, rewritten oregon redistributed without permission.