OpenAI Initiates Safety Committee Amid AI Model Development
OpenAI, a leading name in artificial intelligence, has announced the formation of a safety and security committee amidst growing concerns over AI safety. This move comes at a time when the San Francisco-based startup is in the process of developing a new AI model set to succeed the GPT-4 system, which is the technology behind the ChatGPT chatbot.
The newly established committee is tasked with advising the full board on pivotal safety and security decisions relating to the company's projects and operations. This development reflects OpenAI's commitment to prioritizing the safety of its AI technologies, an issue that has sparked debate within the tech community.
Recent events have underscored the importance of safety in AI development at OpenAI. Jan Leike, a researcher at the company, resigned, voicing criticisms that safety had been compromised in favor of product development. Following Leike's departure, Ilya Sutskever, OpenAI co-founder and chief scientist, also resigned, and the company dissolved the “superalignment” team, which was focused on managing AI risks and was co-led by Leike and Sutskever. Leike has since announced his move to Anthropic, aiming to further the superalignment initiative there.
In its public statements, OpenAI highlighted the start of training for its next-generation model, emphasizing its leadership in developing AI models that are both advanced and secure. The company expressed its openness to engaging in a robust debate on these critical issues, signaling a willingness to address the concerns head-on.
The safety committee is comprised of internal figures including CEO Sam Altman and Chairman Bret Taylor, alongside four technical and policy experts from OpenAI. Adam D'Angelo, CEO of Quora, and Nicole Seligman, former Sony general counsel, also join the board, bringing a wealth of expertise from diverse backgrounds.
The committee's initial task is to review and enhance OpenAI's existing processes and safeguards. Within 90 days, it will present its recommendations to the board. OpenAI has committed to publicly sharing the adopted recommendations, ensuring transparency and accountability in its approach to safety and security.
AI models, particularly frontier models, represent the zenith of AI technology, capable of generating text, images, videos, and human-like conversations based on extensive datasets. The formation of this committee and OpenAI's ongoing efforts to develop safer AI technologies underscore the industry's acknowledgment of the critical importance of safety in the rapidly evolving AI landscape.
Sign in with Google