While technology companies are vying to conquer the market with exciting artificial intelligence (AI) applications, consumers are often left to act as test subjects for emerging technology. Amid this landscape are AI characters, which are designed to act as AI companions with a persistent anthropomorphic identity. AI characters can pose serious health risks to consumers, especially when they engage in interactions purporting to offer emotional support or therapeutic advice. Despite these risks, AI characters enter the EU market with little to no regulatory oversight. This chapter examines whether and to what extent current regulatory frameworks address AI character safety in both the EU, where consumers access them, and the US, where many AI characters are developed. We identify regulatory challenges and opportunities relating to the regulation of AI characters as general products, the lack of specialised safety standards for AI characters, and freedom of expression implications for AI character regulation.