OpenAI Establishes Safety Committee and Begins Training New AI Model
On Tuesday, OpenAI announced the formation of a new safety and security committee and the commencement of training for its latest AI model intended to replace the current GPT-4 system. This development was revealed through a company blog post.
The San Francisco-based artificial intelligence company detailed that the new committee will provide guidance on critical safety and security decisions. Notable members of this committee include OpenAI CEO Sam Altman, Chairman Bret Taylor, Quora CEO Adam D’Angelo, and former Sony general counsel Nicole Seligman. Additionally, four technical and policy experts from OpenAI will join the committee.
Amid ongoing scrutiny about AI safety, these steps come as OpenAI aims to address concerns following the recent resignation of key researchers like Jan Leike. These researchers had criticized the company for prioritizing product features over safety measures.
OpenAI, which is renowned for AI tools such as ChatGPT and DALL-E, has faced legal challenges. Recently, actress Scarlett Johansson raised concerns that a voice used by OpenAI sounded very similar to hers, leading to potential legal action. Additionally, the company is contending with lawsuits from news organizations over alleged copyright infringements during AI training.
Emphasizing their commitment to transparency, OpenAI stated that the committee’s first task will be to review and refine current safety processes. The company plans to publish the committee’s recommendations within 90 days.
Through this step, OpenAI aims to balance innovation with responsible governance as it continues to develop advanced AI technologies.