EU Probes X’s Grok AI Under DSA; Feed Changes Likely for Portuguese Users

The European Commission has launched a fresh formal inquiry into Elon Musk’s X platform after it quietly plugged its Grok artificial-intelligence engine into timelines—a step that could reshape how news, advertising and even deep-fake images reach screens in Portugal.

Why This Matters

• Harsher fines incoming: Under the Digital Services Act (DSA), Brussels can dock up to 6 % of X’s global revenue or suspend risky features.

• Deep-fake crackdown: Investigators are zeroing in on AI-generated sexual or child-abuse imagery, content that Portuguese users have flagged in recent months.

• Timeline shake-ups: If Grok’s recommendation model is found unsafe, X may be forced to switch off or radically alter its feed across the EU, including Portugal.

• Advertiser impact: Lisbon-based brands could face stricter disclosure rules if ad-targeting data flows through Grok.

Brussels Opens a Second Front

The European Commission’s Directorate-General for Communications Networks (DG CONNECT) confirmed it is widening an existing December 2023 probe to specifically inspect Grok. Regulators want to know whether X drafted the mandatory risk-assessment report—a core DSA requirement—before rolling the chatbot-style algorithm out. The investigation will explore how the new AI ranks and amplifies posts, and whether X installed “effective and proportionate” safeguards against illegal content such as manipulated nudes or extremist material.

How Grok Alters Your Feed

Unlike the legacy “For You” timeline, Grok draws on the fire-hose of public posts to generate real-time summaries, punchy jokes and sometimes controversial answers. Because it is both a writing tool and a recommender system, it can instantly fabricate images or text and promote them. Digital-rights advocates in Portugal argue that blurring creation with curation risks turbo-charging misinformation right before the 2026 European elections.

Penalties on the Table

Under the DSA, Very Large Online Platforms—X falls in that category because it tops 45 million monthly EU users—must show they have trimmed systemic risks. Failure can trigger:

Priority corrective orders with strict deadlines.

Recurring fines up to 6 % of annual turnover (roughly €8 billion for X, based on 2025 estimates).

As a nuclear option, a temporary ban on specific services within the Union.

What This Means for Residents

For everyday users in Portugal, the biggest near-term change may be a cleaner, slower feed if X is obliged to pause Grok. Those who rely on X for news could see Portuguese-language content resurface higher in rankings while experimental AI banners disappear. Parents should expect new age-verification prompts when Grok tries to display sensitive images. Creators and SMEs who boost posts will need to label AI material or risk account throttling once updated ad-transparency rules kick in.

Portugal’s Watchdogs Are Already Watching

The Portugal National Cybersecurity Centre (CNCS) and the Comissão Nacional de Proteção de Dados often mirror EU enforcement locally. Officials tell us they stand ready to assist Brussels with on-site inspections or server-log requests if X fails to cooperate. Any EU-level fine automatically compels X to set up a single-point-of-contact in Lisbon for faster takedown orders, shortening response times for abusive content flagged by Portuguese users.

Timeline and Next Steps

Brussels has not published a deadline, but similar DSA cases took 4-9 months. Investigators can still impose interim measures—think forced feature suspensions—while the file is open. X will likely submit its Grok risk report in the coming weeks; insiders say the company is considering a regional-only toggle that lets EU users opt out of AI-driven feeds entirely.

For Portugal’s 2.7 million X account holders, the message is clear: scroll with caution, expect more pop-ups asking for consent, and keep an eye on civic-society hotlines that record AI abuse. Brussels is betting that turning up the regulatory heat now will make future deep-fake crises less likely when Europe heads to the polls next spring.