According to Handelsblatt, the European Commission is preparing to classify ChatGPT as a “very large online search engine” under the DSA. Internal sources cited by the German newspaper suggest a decision in the coming days based on the threshold of 45 million monthly users in the EU as outlined in the regulation for VLOP/VLOSE.
OpenAI has not commented. The Commission indicates that it is reviewing usage data. Being classified as a “very large service” triggers enhanced obligations: assessment and mitigation of systemic risks, increased transparency on algorithms and content, access to data by authorities and researchers, and stricter reporting/recourse mechanisms.
What would change with a “search engine” classification
The DSA distinguishes between platforms and search engines. Applying the “search” regime to a conversational chatbot would require presentation and traceability requirements for results similar to those of Google or Bing: clarity on rankings, identified advertising, and limits on response manipulation. This would push OpenAI to thoroughly document its sources and data pipelines for the EU.
Operationally, this could impact integrations of ChatGPT into browsers and assistants, as well as products that display generative responses instead of links. The tightening would also focus on moderation of outputs for sensitive topics and protection of minors, with regular audits.
A milestone for the regulation of generative AI
If the classification is confirmed, the EU would send a signal: generative AI services on a very large scale will be treated as access points to information subject to transparency and risk management obligations equivalent to search engines. For the ecosystem, this would accelerate the convergence between DSA and the future AI Act, with a significant compliance cost for providers and integrators.
Source: ITHome
