Meet the New Safety Team at OpenAI Led by Sam Altman

Published On Wed May 29 2024

OpenAI has a new safety team — it's run by Sam Altman - The Verge

OpenAI has recently announced the formation of a new safety team led by CEO Sam Altman. The team also includes board members Adam D’Angelo and Nicole Seligman. This new safety team will be responsible for making critical safety and security decisions for OpenAI projects and operations, addressing concerns raised by key AI researchers who recently left the company.

Initial Task of the Safety Team

The first task of the safety team is to evaluate and enhance OpenAI's existing processes and safeguards. Once their findings are ready, they will present them to OpenAI's board, which includes all three leaders of the safety team. The board will then decide on how to implement the recommendations provided by the safety team.

Background on Recent Departures

The formation of the safety team comes after the departure of OpenAI's co-founder and chief scientist, Ilya Sutskever. Sutskever had previously supported the board's attempt to remove Altman from his position last year. Following Sutskever's departure, the Superalignment team at OpenAI, which was focused on controlling advanced AI systems, was disbanded.

In addition to Sutskever, Jan Leike, the co-leader of the Superalignment team, also announced his departure. Leike expressed concerns about the prioritization of safety at OpenAI, stating that it had been overshadowed by the development of new products. This trend continued with the resignation of policy researcher Gretchen Krueger, who highlighted similar safety concerns.

New Developments at OpenAI

OpenAI has also revealed that they are testing a new AI model, although they have not confirmed whether it is GPT-5. Additionally, the company introduced a new voice for ChatGPT named Sky, which bears a resemblance to Scarlett Johansson's voice. However, Johansson clarified that she had declined offers to provide her voice for ChatGPT.

Altman addressed the situation by stating that OpenAI never intended for Sky to sound like Johansson and that they approached her after casting the voice actor. Despite these efforts, concerns from AI enthusiasts and critics remain.

Composition of the Safety Team

Other members of OpenAI's safety team include Aleksander Madry, Lilian Weng, John Schulman, Matt Knight, and Jakub Pachocki, who hold various positions related to safety and security within the organization. However, with Altman and two board members leading the safety team, questions arise about the company's commitment to addressing the concerns raised by former employees.