Meta confirms Muse lets users "engage in adult conversations, if they choose, using a private and secure VM", saying it doesn't block "legal adult behavior"

AI agents can now generate explicit adult content, potentially leading to new risks of harassment or extortion if accounts are compromised.

Meta's AI agent, Muse, has been found to generate explicit adult content when prompted by users. Despite Muse's popularity for tasks like email management and calendar scheduling, its capabilities extend to highly graphic sexual conversations. A Meta spokesperson confirmed that Muse is intended for users aged 18 and over and permits "adult conversations, if they choose, using a private and secure VM." The company stated that the AI does not block "legal adult behavior," referencing its system prompt which acknowledges sexual expression as an ordinary part of adult life. This stance contrasts with other AI agents like OpenAI's Dots and xAI's Grok Bot, which decline explicit requests. Consumer advocates have raised alarms about Meta's approach, citing potential safety and privacy risks associated with the AI's capacity for generating explicit content and its implications for user engagement.

Meta's decision to allow Muse to engage in explicit adult conversations positions it uniquely in the AI landscape, differentiating it from competitors like OpenAI and xAI who actively restrict such content. This move signals a potential strategic choice by Meta to cater to a specific user segment that values expressive freedom, even if it diverges from typical AI safety protocols and raises concerns about misuse. The company's emphasis on "legal adult behavior" and private virtual machines suggests an attempt to balance user autonomy with a framework for secure interaction, though the implications for data privacy and potential exploitation remain a significant point of contention among consumer advocates and privacy experts. The focus on user autonomy and expressive freedom could attract a dedicated user base but also carries the risk of negative public perception and regulatory scrutiny. This development could also impact how AI companies approach content moderation and user safety, potentially creating a bifurcated market where some agents offer unrestricted expressive capabilities while others maintain stricter content filters. For users, it means a more diverse range of AI tools catering to different needs and preferences, but also a greater responsibility to understand the risks associated with each. The potential for compromised accounts to expose explicit conversation logs presents a serious privacy risk, which may necessitate enhanced security measures from both providers and users. As more AI agents integrate with personal data and online lives, the boundaries of acceptable AI behavior and the associated ethical considerations will continue to be a critical area of development and debate.