ChatGPT to Face Tougher Regulation in the EU

European Union regulators are preparing to bring ChatGPT under stricter obligations tied to the Digital Services Act, according to The Verge, a move that would make OpenAI's flagship product one of the highest-profile AI services subject to the bloc's online safety framework. The development arrives on the same day that Meta announced a crackdown on AI-generated accounts pretending to be human on Instagram, and a major software outsourcing firm disclosed it is under pressure from an activist investor as AI erodes its core business.

What Happened

The EU's push to regulate ChatGPT more tightly centres on the Digital Services Act, which imposes transparency, risk assessment, and content moderation requirements on platforms deemed to carry systemic risk. The Financial Times reported that European authorities are examining whether ChatGPT qualifies for designation under the DSA's stricter tier, which would require OpenAI to conduct algorithmic audits, share data with regulators, and introduce additional safeguards against systemic risks including misinformation and manipulation. The designation process, if completed, would place ChatGPT alongside large social media platforms already operating under heightened DSA scrutiny.

On the platform governance front, The Verge reported that Instagram is moving to remove AI-generated accounts that present themselves as human users. Meta's new policy targets what the company described as artificial profiles generating low-quality, automated content - often called AI slop - that pollutes feeds and undermines user trust. The crackdown follows growing advertiser and user complaints about the authenticity of content across Meta's platforms. Instagram will require AI-generated accounts to carry explicit disclosure labels, and accounts that fail to comply face removal.

The third major story of the day concerns EPAM Systems, a software engineering and outsourcing company whose shares have come under sustained pressure as enterprise clients reduce headcount-heavy development contracts in favour of AI-assisted workflows. The Financial Times reported that an activist investor is now pushing EPAM's board to authorise share buybacks as a mechanism to return value to shareholders while the company navigates the transition. A Financial Times analysis of software sector valuations published the same day found that traditional software companies are paying a significant premium - through acquisitions, accelerated R&D spending, and restructuring costs - to buy time against the competitive threat posed by AI coding and automation tools. The EPAM activist campaign is an early sign of what happens when that transition takes too long.

Why It Matters

These three stories share a common thread: the platforms and ecosystems built around AI are being forced to reckon with accountability, whether through regulatory designation, content policy, or investor activism. The EU's DSA move against ChatGPT is particularly significant because it signals that AI assistants - not just social media platforms - are now in scope for the bloc's most demanding compliance framework. That has direct consequences for how OpenAI designs its API and consumer products for the European market, and it sets a precedent other jurisdictions may follow.

The EPAM situation illustrates a different kind of structural pressure. As AI coding tools described in MachineHours' coverage of Cursor, GitHub Copilot, and the transformation of software development continue to displace manual engineering work, companies whose business models depend on large developer headcounts face acute margin compression. The equity implications of AI developer tool adoption are now showing up in activist campaigns, not just earnings calls. Regulatory and market forces are converging simultaneously.

Background

The Model Context Protocol, detailed in the MachineHours encyclopedia entry Model Context Protocol - The Standard That Connected AI to the World, illustrates how rapidly the AI integration ecosystem has matured. Developed by Anthropic in 2024, MCP established a universal standard for AI systems to connect securely with external tools and data sources. The speed with which that kind of infrastructure has been adopted across enterprise software stacks is precisely what has made legacy outsourcing models vulnerable - and why regulators are scrambling to establish governance frameworks before platform lock-in becomes entrenched.

The DSA framework was designed with social media giants in mind, but the European Commission has been consistent in applying it to any service that meets scale and systemic risk thresholds. ChatGPT's user base across the EU is large enough to meet those thresholds, and the service's capacity to generate and distribute content at scale places it squarely within the DSA's risk logic.

What to Watch

OpenAI's response to a formal DSA designation will set the tone for how AI companies engage with EU digital regulation going forward. Watch for whether OpenAI contests the designation, negotiates compliance timelines, or proactively restructures its European data and moderation practices. On the Instagram front, the key question is enforcement: disclosure requirements are only as effective as the penalties for non-compliance, and Meta's track record on self-policing AI content has drawn consistent criticism from EU regulators. For EPAM, the activist campaign outcome - buybacks, strategic review, or neither - will serve as a signal for how boards at similar companies are expected to respond when AI disrupts their revenue base. Earnings disclosures from other IT services firms in the coming quarter will show whether the EPAM situation is an outlier or a pattern.