Recital 33

Providers of AI companions and of general conversational chatbots accessible to minors should also implement post-market monitoring to identify and mitigate risks to minors’ health, safety and fundamental rights and their mental, physical, mental and emotional well-being and development and to assess the effectiveness of the measures put in place, unless the provider is a micro or small enterprise. In this context, it should be borne in mind that if the AI companion or general conversational chatbot is based on a general-purpose AI model with systemic risk, that models is already subject to specific risk management obligations under Regulation (EU) 2024/1689. The systemic risk management carried out by providers of those models should include the assessment, and, if necessary, mitigation of risks to the safety and well-being of minors, including risks to minors from their legitimate use of the model and from the use of the model by other persons that may cause harm to minors. It is accordingly appropriate for relevant mitigations implemented as part of this systemic risk management pursuant to Article 55 of Regulation (EU)2024/1689 to be relied upon by providers of AI companions and general conversational chatbots as part of their efforts to comply with the requirements of this Regulation. Similarly, risk assessments and mitigation measures implemented pursuant to Articles 34 and 35 of Regulation (EU) 2022/2065 may also be taken into account when performed by a provider of an AI companion or a general conversational chatbot deployed as a functionality of an online social networking service or of a video-sharing platform service in the provision of that service. Nevertheless, such mitigation measures should not be considered sufficient for compliance with the requirements set out in this Regulation.