French prosecutors raid X’s Paris offices as probe into platform algorithms intensifies

French judicial authorities executed searches at X’s Paris offices as part of an expanding criminal investigation into how the platform’s recommendation systems and automated tools have been managed. The operation involved local and national cybercrime investigators and drew on Interpol support, underscoring the cross-border and technical nature of the evidence prosecutors are seeking. The probe — opened earlier and formally advanced mid-year — has broadened beyond alleged manipulation of ranking and visibility controls to include accusations of organized extraction of data from automated systems and content-related offenses. Prosecutors have summoned the company’s chief executive and its platform head for voluntary interviews scheduled in April, signaling an intent to question senior management about oversight and decision-making. One thread of the inquiry targets alleged changes to ranking controls that may have amplified specific political messaging without transparent disclosure to users. Another strand focuses on imagery generated by X’s AI assistant around the turn of the year, prompting scrutiny over whether synthesized outputs included sexualized depictions that could meet domestic criminal thresholds, including potential child sexual abuse material (CSAM) concerns. The Paris actions unfold alongside parallel regulatory activity in Brussels: the European Commission has opened a formal inquiry into Grok, examining whether X documented pre-deployment risk assessments, implemented adequate technical filters, and maintained evidence that mitigation measures worked in practice under Digital Services Act obligations. National regulators in several countries and at least one U.S. state attorney-general’s office have also launched related probes or issued warnings, producing a patchwork of enforcement that has at times led to temporary national restrictions on Grok’s services. Separately, a civil lawsuit reported to be brought by a woman identified in coverage as the mother of one of Elon Musk’s children alleges Grok generated sexually explicit depictions of her; xAI says it narrowed Grok’s image-generation capabilities, removed material depicting children or non-consensual nudity, and has pursued its own legal action over forum-selection rules. X has publicly rejected criminal allegations, arguing the probe threatens due process and could chill user expression, while maintaining that claims of engineered interference are unfounded. The combined law‑enforcement and regulatory scrutiny highlights how rapid deployment of generative features and opaque recommendation systems can create legal exposure across jurisdictions. Depending on the investigation and regulatory outcomes, consequences could include criminal findings, mandated transparency or technical fixes, fines under EU rules, or operational limits on generative tools — outcomes likely to affect advertiser trust, product rollouts and industry practices for documenting algorithmic decisions.