Today, the European Commission officially classified ChatGPT as a Very Large Online Search Engine, while Reddit and Roblox received the designation of Very Large Online Platforms under the auspices of the Digital Services Act (DSA).
All three entities must adhere to the stipulated regulations by the close of December 2026. Notably, ChatGPT marks the inaugural inclusion of a generative AI chatbot in this category.
Understanding the Digital Services Act (DSA)
The Digital Services Act is a legislative framework in the EU, designed to delineate the obligations of online services for managing illegal and detrimental content.
Enacted in November 2022, the DSA has progressively applied to all online services within the bloc since February 2024.
The DSA operates on a bifurcated system. The foundational tier is universally applicable: it mandates the publication of comprehensible terms, provides users with mechanisms to report illicit content, demands transparency regarding content removal, and ensures avenues for appeals. Conversely, the elevated tier, which these three entities now occupy, imposes significantly more stringent requirements.
Criteria for Ascending to the Upper Tier
Ascension to this upper tier is contingent upon statistical metrics. Each platform and search engine engaging with users in the EU is required to disclose its average monthly user count, updating these figures biannually. When a service exceeds 45 million users, the Commission retains the authority to confer designation.
OpenAI reported approximately 159 million monthly users utilizing ChatGPT’s search capabilities over the six months leading up to March 2026.
Reddit indicated 57.2 million users, while Roblox estimated about 48 million. This designation expands the cohort of companies under the rigorous EU regulations to a total of 28, encompassing prominent platforms such as Facebook, Instagram, TikTok, YouTube, X, Amazon, Google Search, Bing, and Wikipedia.
A Very Large Online Platform is defined as one that hosts user-generated content, exemplified by Reddit and Roblox, whereas a Very Large Online Search Engine processes inquiries and seeks out answers, as determined by the Commission’s designation of ChatGPT.
Obligations for the Newly Designated Services
By December’s end, each company is required to furnish a comprehensive written evaluation of the systemic risks inherent in their respective products.
This is not a promotional document; it must candidly address how their services could propagate illicit content, endanger children, adversely impact users’ mental or physical well-being, infringe on fundamental rights, distort electoral processes, or jeopardize public safety.
Subsequently, companies must implement corrective measures based on their findings. The law permits revisions in product design or recommendation algorithms, and mandates the establishment of an internal compliance unit to oversee these actions.
Once annually, an independent audit will scrutinize this entire process, with the requirement for companies to respond to the auditor’s recommendations.
Additionally, each organization must maintain a public, searchable archive of all advertisements it hosts, detailing who funded them and their target demographics, while also granting vetted academic researchers access to pertinent data to evaluate potential harms independently, rather than relying solely on the companies’ assertions. Furthermore, at least one unfiltered content feed must be provided.
The Complexity Surrounding ChatGPT’s Designation
The designation of Reddit and Roblox was anticipated, as both platforms are characterized by user-generated posts.
ChatGPT’s classification, however, ignited debate. The DSA framework was established prior to the launch of ChatGPT, which complicates its categorization.
The legislation addresses platforms and search engines, yet a chatbot fits neither distinctly. The Commission opted not to defer until the AI Act, instead categorizing ChatGPT as a hybrid service and designating it as a search engine due to its capacity to respond to user inquiries, often by sourcing information from the web.
Unlike traditional search engines that present multiple links while leaving evaluation to the user, a chatbot provides single, assertive answers, often integrating sources that may be obscured or absent.
Here, the evaluative responsibility shifts from the user to the AI model itself. Thus, the inherent risks associated with singular answers necessitate comprehensive scrutiny.
Nevertheless, this DSA designation does not impose restrictions on what OpenAI and ChatGPT may articulate; instead, it compels the organization to investigate the societal impact of its offerings, document findings, rectify identified issues, and transparently disclose the methodologies employed—an entirely reasonable expectation for a service engaging with 159 million users.
Challenges Confronting Roblox
Roblox presents unique challenges, primarily as its user base consists largely of minors. As of July 2026, 170 lawsuits were consolidated within a US federal case concerning child exploitation on the platform, with Roblox recently reaching a settlement with Alabama amounting to $12.2 million.
In response, the company has implemented facial age estimation technology and categorized users into age groups before granting chatting privileges. The EU classification permits external evaluation of the efficacy of these measures.
Conversely, Reddit’s designation stems from its reliance on volunteer moderators and its position as a foundational layer for AI-generated responses, which are licensed to OpenAI and Google for that very purpose.
Imposing Sanctions
Failure to comply with the EU’s designation could incur penalties of up to 6% of an entity’s global annual revenue.
The Commission has previously indicated its readiness to enforce such penalties, as demonstrated when X (formerly Twitter) faced a fine of €120 million on December 5, 2025, regarding their visibility filtering practices.
Global Implications
Though these audits are confined to the EU, the underlying concerns are universally relevant. The financial investment in safety engineering is substantial, and replicating efforts across regions proves inefficient.
Consequently, it is likely that many of these innovations will extend beyond Europe. For instance, Roblox has globally deployed its age verification protocols rather than confining them to specific territories.
Source link: Tech-ish.com.
