Meta is dissolving its central Responsible AI (RAI) team, a specialized group focused on addressing ethical and safety concerns in artificial intelligence. The company confirmed that the majority of the team’s members will be reassigned and embedded directly within its primary generative AI product team and AI research divisions.
This strategic shift marks a significant change in how the tech giant approaches AI safety. Rather than having a separate, centralized body for oversight, Meta is moving to integrate safety experts directly into the development lifecycle of its products, including its Llama family of large language models. According to a company spokesperson, this move allows Meta to more effectively scale its safety efforts by having experts work alongside developers from the project outset.
The RAI team was instrumental in creating policies and frameworks to mitigate risks such as algorithmic bias, misinformation, and other potential harms stemming from AI. Its dissolution comes at a critical time when Meta is accelerating its push into generative AI to compete with rivals like OpenAI, Google, and Anthropic. The company argues that embedding these experts will make safety a core component of product architecture, not an external checkpoint.
However, the move has drawn scrutiny from some industry observers. Critics express concern that without a centralized and independent team, safety priorities could be deprioritized in favor of faster product rollouts and commercial objectives. The risk is that embedded experts may lack the autonomy to challenge product decisions or raise red flags effectively. As AI models become increasingly powerful and integrated into society, Meta’s new decentralized approach to safety will be closely watched by regulators and the public alike to see if it strengthens or weakens the company’s commitment to responsible innovation.
English 
Deutsch