TLDR: The European Union is evaluating whether OpenAI’s ChatGPT, specifically its search functionality, should be designated as a ‘Very Large Online Search Engine’ (VLOSE) under the Digital Services Act (DSA). This consideration follows OpenAI’s report of 120.4 million monthly active users for ChatGPT’s search feature in the EU, significantly surpassing the DSA’s 45 million-user threshold for stricter regulations. Such a classification would impose new obligations on OpenAI, setting a significant precedent for the regulation of generative AI across the bloc.
The European Commission is currently assessing whether OpenAI’s ChatGPT, particularly its integrated search feature, qualifies as a ‘Very Large Online Search Engine’ (VLOSE) under the stringent rules of the Digital Services Act (DSA). This potential designation stems from OpenAI’s recent disclosure that ChatGPT’s search functionality recorded an average of 120.4 million monthly users within the EU over the past six months, far exceeding the 45 million-user benchmark that triggers enhanced regulatory oversight for online platforms and search engines.
Should ChatGPT’s search feature be formally designated as a VLOSE, it would mark a pivotal moment in the regulation of generative AI chatbots in Europe. The DSA, designed to create a safer digital environment and ensure a level playing field, imposes a strict due-diligence regime on platforms deemed to pose systemic risks to society. For OpenAI, this would translate into new obligations, including systemic risk and audit reporting, as well as increased transparency requirements.
A spokesperson for the European Commission confirmed that regulators are “currently assessing the information and whether further clarifications are needed” to determine if ChatGPT meets the criteria for designation. A key challenge in this assessment is whether to treat ChatGPT’s search capability as a distinct service or as an inseparable part of the broader chatbot offering. The DSA’s existing categories for intermediary services, such as search engines, marketplaces, and social networks, do not perfectly align with the multifaceted nature of AI chatbots.
Experts note the complexity of this classification. Joan Barata, a Visiting Professor at the School of Law at Católica, expressed some surprise at the potential designation, stating, “It would raise questions about transparency, where privacy concerns might be particularly relevant, and about systemic risk requirements, with clear overlaps between the DSA and the AI Act.” He further elaborated that if chatbots are deployed as specific services by platforms, the designation could extend to the platforms embedding them.
If designated, OpenAI would be required to conduct annual risk assessments, which would also be triggered each time a new functionality with a “critical impact” on systemic risks is deployed. This includes evaluating potential negative impacts on civic discourse, electoral processes, and the protection and mental health of minors. As part of mitigation measures, OpenAI might need to adapt ChatGPT’s systems, design, features, and functioning based on these assessed risks.
Also Read:
- Microsoft and OpenAI Forge Restructured Partnership, Valuing AI Leader at $500 Billion
- OpenAI Offers One-Year Free ChatGPT Go Subscription to All Users in India
OpenAI has already taken steps to comply with DSA obligations, publishing its monthly active recipient figures for ChatGPT search and establishing designated points of contact for both authorities and users in the EU. This proactive approach underscores the company’s recognition of the evolving regulatory landscape for AI technologies within the European Union.
