In response to ChatGPT, the EU is now regulating artificial intelligence not only as a technology, but as a new gateway to online information. (Photo: Shutterstock)

In response to ChatGPT, the EU is now regulating artificial intelligence not only as a technology, but as a new gateway to online information. (Photo: Shutterstock)

The European Commission has classified ChatGPT as one of the ‘very large online search engines’. This classification marks a new stage in the recognition of the role of artificial intelligence in access to information. Reddit and Roblox are also subject to enhanced monitoring.

Is ChatGPT a search engine? As far as the European Commission is concerned, the answer is now yes – at least under European regulations. Brussels has therefore just designated ChatGPT as a ‘very large online search engine’ (VLOSE) within the meaning of the Digital Services Act (DSA). The European regulation thus places OpenAI’s artificial intelligence service in the same regulatory category as major search engines, notably due to its ability to provide results sourced from the web. The decision is far from insignificant. It means that the European Union now regards ChatGPT not only as a conversational artificial intelligence tool, but also as a service that can act as an intermediary in Europeans’ access to online information.

A qualification arising from his new role

The Commission bases its assessment, in particular, on the development of ChatGPT and the integration of internet search functions. The service can now respond to user queries by retrieving information from the web and presenting it in a summarised form. It is this combination that has led Brussels to classify ChatGPT as a ‘hybrid service’ and to subject it to the rules applicable to very large online search engines.

This change is significant because the DSA sets out specific obligations for services that reach a particularly large audience within the European Union, namely an average of 45 million monthly users.

More responsibilities

This new classification subjects ChatGPT to an additional level of obligations. In particular, the service will be required to identify and assess the systemic risks associated with its operation and take measures to mitigate them. The Commission will focus in particular on risks to fundamental rights, public safety, electoral processes and the dissemination of illegal content. The protection of minors, as well as risks to their physical and mental wellbeing, are also among the areas under scrutiny.

Recommendation systems will also require particular attention. In particular, very large digital services must offer users the option of receiving recommendations that are not based on profiling. ChatGPT will also have to undergo regular independent audits and, under certain conditions, allow accredited researchers access to its data. The Commission will also be able to exercise its supervisory powers directly.

AI: the gateway to the web

The European decision comes at a time when the line between traditional search engines and artificial intelligence assistants is becoming increasingly blurred. For years, searching for information online meant entering a query into a search engine such as Google or Bing and browsing through a list of links. AI assistants now offer a different approach: the user asks a question in natural language and receives an answer straight away, sometimes accompanied by sources from the web.

By classifying ChatGPT as a very large search engine, Brussels is therefore formally recognising this shift in usage patterns. The issue is no longer simply how AI generates its responses, but also what role it plays in access to information and the circulation of content online.

Reddit and Roblox under increased scrutiny

ChatGPT is not, in fact, the only service affected by the tightening of European regulation. Reddit – a community-based discussion and social news site – and Roblox – an online video game platform and game creation system – have also been designated as very large online platforms (VLOPs). All three were already subject to the general obligations under the DSA, but their new classification brings with it additional requirements. They have four months to comply with these new obligations.

According to Henna Virkkunen, Executive Vice-President of the Commission responsible for technological sovereignty, security and democracy, these appointments reflect a commitment to strengthening the accountability of digital services that reach a particularly large audience within the EU.

With ChatGPT, however, this responsibility takes on a new dimension: the European Union is now beginning to regulate artificial intelligence not only as a technology, but as a new gateway to online information.