OpenAI Forms Safety and Security Committee to Guide Development of New AI Model

OpenAI has announced the formation of a new Safety and Security Committee, led by board members Bret Taylor, Adam D’Angelo, Nicole Seligman, and Sam Altman. This committee will play a crucial role in steering the development of OpenAI’s next frontier AI model. The committee’s first task is to evaluate and enhance OpenAI’s processes and safeguards over the next 90 days. At the conclusion of this period, they will share their recommendations with the full Board of Directors. It is expected that OpenAI’s new AI model, whether it be called GPT-5 or something else, will not be released for at least 90 days to allow for the evaluation and implementation of safety measures. The recommendations from the Safety and Security Committee are anticipated to be received by the board by August 26, 2024.

The decision to have a 90-day evaluation period is a commonly used timeframe in business to assess and provide feedback on processes. While some have criticized the new committee for being comprised entirely of OpenAI executives, it is important to note that this approach is not uncommon in the tech industry. The previous board of directors at OpenAI faced controversy when they fired CEO Sam Altman, only to reinstate him later due to employee revolt and pushback from investors like Microsoft. The addition of new board members earlier this year aimed to diversify the board’s composition.

OpenAI has recently faced public relations challenges since the release of its latest AI model, GPT-4o. Scarlett Johansson accused the company of approaching her about voicing its new assistant, resembling her character in the movie Her, but OpenAI clarified that they commissioned a separate voice actor and did not intend to imitate Johansson. Additionally, OpenAI’s chief scientist and co-founder Ilya Sutskever resigned, criticizing the company for prioritizing “shiny products” over safety. Despite these challenges, OpenAI has seen success in partnering with mainstream media for training data and has garnered interest from musicians and filmmakers for its Sora video generation model.

OpenAI’s new Safety and Security Committee aims to address concerns raised by critics, regulators, and potential business partners. By focusing on evaluating and enhancing safety measures, OpenAI hopes to build trust and ensure ethical compliance in its AI models. The committee’s recommendations will be crucial in shaping the future of OpenAI’s AI development and will be eagerly awaited by stakeholders in the field.