The European Commission is set to designate OpenAI’s ChatGPT and gaming platform Roblox as “Very Large Online Platforms” under the Digital Services Act, subjecting both to the bloc’s strictest content moderation and transparency requirements.
The designation, expected as soon as August, follows both services surpassing the DSA’s 45 million monthly active user threshold in the EU. ChatGPT’s search functionality alone reported more than 120 million monthly EU users.
What is the Digital Services Act? The DSA is the EU’s flagship content regulation law, in force since 2022. It requires online platforms to clamp down on illegal and harmful content, file transparency reports, detail risk mitigation plans, and pay an annual fee to the European Commission. Companies that breach the DSA face fines of up to 6% of their annual global sales.
🔍 THE BOTTOM LINE
The EU is bringing AI chatbots into the same regulatory framework that governs social media platforms. For OpenAI, this means ChatGPT is no longer just a product — it’s a regulated platform with obligations around content moderation, algorithmic transparency, and child safety. The designation blurs the line between AI tool and media platform, and sets a precedent that could ripple to other AI services.
What the VLOP Designation Requires
Once designated, companies have four months to comply with a suite of obligations that go well beyond what standard AI regulations like the EU AI Act demand:
- Transparency reports: detailed disclosures about content moderation decisions, removals, and algorithmic recommendations
- Risk mitigation plans: systematic assessments of how the platform could be misused, with concrete steps to reduce those risks
- Contact channels: clear reporting mechanisms for both regulators and users, including for suspected criminal activity
- Advertising transparency: if the platform runs ads, it must disclose who paid, how much, and who was targeted
- Annual fee: a contribution to the Commission’s oversight costs, scaled to the platform’s size
For ChatGPT, the advertising requirement may not apply directly, but the content moderation and risk assessment obligations do. OpenAI will need to demonstrate how it handles hallucinated content, disinformation prompts, and potential misuse for generating illegal material.
Why Roblox, and Why Now
Roblox’s inclusion is driven by child safety concerns. The platform has faced criticism over its protections for young users, and recently implemented stronger controls over who children can talk to and what games they can access. The DSA’s emphasis on protecting minors online makes Roblox a natural target.
The platform is also moving into advertising, which triggers the DSA’s transparency requirements around marketing to children. Brussels has been signalling for months that gaming platforms with large underage user bases would face scrutiny.
The Trump Administration Factor
The DSA has become a flashpoint in US-EU trade relations. The Trump administration has accused the EU of censoring American companies, and the designation of ChatGPT — arguably the most prominent US AI product — will intensify that dispute.
The EU’s response is that the DSA applies to all platforms operating in the bloc, regardless of origin. Alibaba’s e-commerce service was fined €550 million earlier in July for failing to properly evaluate unsafe products. In December, X was fined €120 million for deceptive design and lack of transparency. Both companies appealed.
The precedent is clear: the DSA is not selective enforcement. Whether US political pressure can change that calculation is an open question.
What This Means for AI Companies
The designation is significant because it treats an AI chatbot as a content platform rather than a software tool. Under the EU AI Act, ChatGPT is regulated as a general-purpose AI model. Under the DSA, it would be regulated as a platform that disseminates content to 120 million EU users.
This dual classification creates a new compliance layer. OpenAI will need to satisfy both frameworks simultaneously — the AI Act’s requirements around model documentation and risk management, and the DSA’s requirements around content moderation and transparency.
Other AI services with large EU user bases — Google’s Gemini, Anthropic’s Claude, Meta’s AI assistant — could face similar designation if they cross the 45 million user threshold. The Commission has said decisions are taken “on a case-by-case basis.”
NZ Angle
New Zealand’s own digital regulation framework is still developing. The EU’s move to classify AI chatbots as content platforms raises a question for Wellington: should services like ChatGPT be subject to NZ’s content moderation expectations? The Department of Internal Affairs has been consulting on online content regulation, but no AI-specific platform rules exist yet.
For NZ companies building AI tools with large user bases, the EU designation sets a regulatory template that could inform domestic policy. The compliance costs of dual AI Act and DSA obligations are not trivial — and they fall hardest on smaller companies that lack Meta’s or Google’s legal teams.
❓ FAQ
What happens if OpenAI doesn’t comply with the DSA? Fines can reach 6% of annual global revenue. For OpenAI, which reportedly generated $3.4 billion in revenue last year, that could mean hundreds of millions. The Commission can also require changes to how the service operates in the EU.
Does this affect ChatGPT users in New Zealand? Not directly. The DSA applies to services offered in the EU. However, companies often implement compliance changes globally rather than maintaining separate versions, so NZ users may see new transparency features or content moderation tools as a side effect.
Why is ChatGPT being treated as a platform rather than a tool? Because it generates and disseminates content to 120 million EU users. The EU’s view is that at that scale, the service functions as a content distribution platform, not just a software utility — especially given concerns about AI-generated disinformation.
What’s the difference between the DSA and the EU AI Act? The DSA regulates platforms based on their size and content dissemination. The AI Act regulates AI systems based on their risk level. ChatGPT could now be subject to both: the AI Act as a general-purpose model, and the DSA as a very large online platform.
🔍 THE BOTTOM LINE
The EU is drawing a line: AI chatbots at scale are content platforms, not just software. For OpenAI, the VLOP designation adds a compliance burden that social media companies have been navigating for years — content moderation infrastructure, transparency reporting, and regulatory fees. For the AI industry, it’s a signal that the regulatory perimeter is expanding beyond model safety into platform accountability. The companies that build compliance into their architecture will adapt; those that treat it as an afterthought will learn what 6% of global revenue feels like.
📰 Sources
- Bloomberg — ChatGPT, Roblox to Fall Under Strictest EU Rules for Platforms
- Straits Times — ChatGPT, Roblox to fall under strictest EU rules for platforms
- Computing.co.uk — EU set to classify ChatGPT under strict online platform rules
- European Commission — DSA: Very large online platforms and search engines
