OpenAI has dismantled its dedicated preparedness team, according to reporting from the Financial Times. The unit had been responsible for identifying potential risks posed by AI models and developing strategies to address those dangers. Going forward, safety-related responsibilities will be distributed across different departments focusing on specific risk areas, including biosecurity and cybersecurity threats.
The dissolution of this team represents another significant reorganization at the artificial intelligence company. Over recent years, OpenAI has scaled back several safety-focused initiatives, including its superalignment and AGI readiness programs. Several key personnel have recently departed the company, including ethics lead Chloé Bakalar and head of safety Johannes Heidecke, among others.
The timing of these changes coincides with OpenAI’s anticipated initial public offering, which is expected to be substantial. Dylan Scandinaro, who previously led the preparedness team and was hired from competitor Anthropic earlier this year, will now concentrate on examining the implications of self-improving artificial intelligence systems.
Critics have raised concerns about the company’s strategic priorities, with some former employees arguing that OpenAI is prioritizing product development over safety considerations during this period of significant organizational change.
Originally reported by Theverge →
About the Author
