EU Just Made ChatGPT a Target — and That Changes Everything

The Box Has Been Drawn

The European Commission has reclassified ChatGPT as a Very Large Online Search Engine under the Digital Services Act. That might sound like bureaucratic wordplay — but it is one of the most consequential boundary decisions in AI regulation to date. For the first time, a generative AI system has been folded into a legal framework originally designed for social media and search platforms.

It also means ChatGPT now faces the same scrutiny as Reddit and Roblox, both named at the same time. The three entities share one thing: each has more than 45 million average monthly users in the EU. Once you cross that threshold, the DSA stops treating you as a neutral tool and starts treating you as a public square with responsibilities.

That distinction matters enormously for OpenAI.

What the DSA Actually Requires

The Digital Services Act, which came fully into force in 2024, was built to rein in how platforms moderate content and target users. The new designations obligate ChatGPT to comply with several specific provisions:

• Platforms cannot target ads to minors.
• Ad targeting based on sexual orientation, religion, ethnicity, or political belief is prohibited.
• Recommendation algorithms must be disclosed with greater transparency.
• Platforms must demonstrate they are mitigating systemic risks to user mental health and the spread of illegal content.

ChatGPT has until the end of December 2026 to comply. That is a narrow window for a company that has never built a compliance apparatus for European content-moderation law.

Henna Virkkunen, the EU executive vice-president overseeing tech sovereignty, called the move necessary because these platforms now exert “a large impact on our citizens and society.” The phrasing is deliberate. The Commission is no longer asking whether generative AI deserves oversight. It is asking how aggressively.

Why This Is Different From Previous AI Regulation

The EU has already passed the AI Act, a comprehensive piece of legislation that categorizes AI systems by risk level. Critics have noted that the AI Act focuses primarily on the models themselves — their training data, safety evaluations, and high-risk use cases. The DSA, by contrast, focuses on how those models are delivered to users and what harm they cause in practice.

Reclassifying ChatGPT under the DSA effectively creates a dual-regulation layer. OpenAI must now satisfy the AI Act’s model-level requirements and the DSA’s platform-level obligations simultaneously. That overlap is untested. There is no precedent for how a generative AI chatbot answers to both frameworks, especially when their definitions of harm and accountability diverge.

The AI Act regulates what the model can do. The DSA regulates what the platform allows to happen around the model. When ChatGPT generates harmful content, the AI Act looks at the training and the architecture. The DSA looks at the distribution, the user interface, and the moderation choices. Both apply at once.

Who Wins and Who Loses

European regulators win the most. They have extended their jurisdictional reach into a domain that was previously hard to pin down. Generative AI did not fit neatly into existing platform laws. The DSA reclassification closes that gap without requiring new legislation.

Consumer-advocacy groups also win. The rules around minors, mental health, and algorithmic transparency are exactly the protections these organizations have lobbied for. ChatGPT’s interaction patterns — personal, persistent, often used by young people — make those protections especially relevant.

OpenAI loses something intangible but significant: speed. The company’s product roadmap has always been defined by how quickly it can ship. A compliance deadline of December 2026 forces the company to slow down in Europe at a moment when its competitors are accelerating. Every feature launch now carries a new question — does this comply with the DSA? Has the risk assessment been completed? Is the moderation pipeline ready?

Tighter constraints on ad targeting also hit revenue. OpenAI’s free tier has always depended on optional upgrades and partnerships. If the DSA restricts how OpenAI can monetize its European user base, that revenue stream shrinks. That is a direct business impact, not a theoretical one.

What This Means for OpenAI’s Product Roadmap

OpenAI’s product strategy has relied on a single interface delivering multiple capabilities. ChatGPT is both a search engine and a content generator. The DSA now treats it as both a search engine and a platform. That forces OpenAI to build separate compliance infrastructure for two legal identities it previously occupied as one.

The company will likely need to invest in content-moderation teams, algorithmic-transparency reporting, and risk-assessment processes tailored to the European market. Those are not one-time costs. They are ongoing operational requirements that scale with user growth. OpenAI will face a structural cost increase that its competitors in the US and China do not share.

This could also affect how OpenAI designs features for Europe. The DSA’s restrictions on targeted advertising and algorithmic recommendation mean the company may have to offer European users a less personalized experience. That is a product decision driven by regulation, not by engineering or market research. The result could be a European version of ChatGPT that is fundamentally different from the global product — slower recommendations, fewer tailored features, and a more limited advertising ecosystem.

The Global Ripple Effect

The non-obvious implication of this ruling is not what happens in Europe. It is what happens everywhere else.

When the EU classifies a US tech company under its Digital Services Act, it sets a regulatory precedent that other jurisdictions will cite. Brazil, India, Japan, and South Korea have all signaled interest in their own platform-regulation frameworks. They do not need to draft entirely new laws. They can point to the EU’s classification of ChatGPT and adopt similar language.

This is the Brussels Effect in action — the phenomenon where EU regulations become de facto global standards because multinational companies prefer to comply once rather than navigate dozens of conflicting rules. But it works in reverse too: when the EU stretches its laws to cover new categories like generative AI, it teaches other governments how to do the same.

For OpenAI, the long-term risk is regulatory fragmentation. Instead of one AI Act and one DSA, the company may face a patchwork of national laws that each treat generative AI differently. Some countries will classify it as a tool. Others will classify it as a platform. Still others will create entirely new categories. The compliance burden multiplies with every jurisdiction that adopts the EU’s logic.

What Comes Next

The December 2026 deadline is the first real test. OpenAI must demonstrate compliance with DSA requirements or face enforcement action. The Commission has the authority to impose fines and, in extreme cases, suspend services.

Beyond that, the legal architecture is unclear. No court has yet ruled on how the DSA applies to generative AI. OpenAI will almost certainly challenge aspects of the classification. It may argue that ChatGPT is a tool, not a platform. It may argue that the 45-million-user threshold should not apply to search-engine-style AI. The outcome of those disputes will define the regulatory landscape for the entire industry.

What is already certain is that the era of regulatory ambiguity for generative AI is over. The EU has drawn a line. The question now is whether OpenAI and other companies can adapt quickly enough to stay on the right side of it.