OpenAI establishes safety committee amid concerns over AI risks

OpenAI announces the establishment of a new safety committee to evaluate the company’s protocols and protections for advanced AI research

Following recent departures and questions about its commitment to AI safety, OpenAI announced the establishment of a new safety committee on Tuesday. This move comes as the business prepares to construct its next-generation AI model, which is projected to outperform GPT-4, the system that drives ChatGPT.

The group, which will be led by CEO Sam Altman, will include both board members and executives. Over the next 90 days, it will thoroughly evaluate OpenAI’s current protocols and protections for advanced AI research.

Throughout this important moment, OpenAI has highlighted its willingness to engage in discourse and debate. “While we are proud to build and release industry-leading models on both capabilities and safety, we welcome a robust debate at this important juncture,” the organization said.

As part of the evaluation process, OpenAI will solicit feedback from other experts such as former US cybersecurity officials Rob Joyce and John Carlin. Joyce, who formerly directed activities at the National Security Agency, and Carlin, a former top Justice Department official, will provide insight into the committee’s discussions.

The committee’s responsibilities include reviewing current AI safety standards and making recommendations for prospective improvements or additions. After the 90-day assessment, the results will be given to the whole OpenAI board before being made public.

The formation of the safety committee comes after OpenAI’s “superalignment” team was disbanded earlier this month, raising questions about the company’s goals in terms of AI safety. In tweets on X, formerly known as Twitter, Jan Leike, co-leader of the disbanded team, accused OpenAI of prioritizing “shiny new products” above crucial safety work.

In addition to internal issues, OpenAI has encountered external criticism, such as the creation of an AI voice that some think closely resembles actress Scarlett Johansson. Despite the company’s protestations, the episode sparked larger arguments about the ethical implications of artificial intelligence.

With the introduction of the safety committee, OpenAI hopes to address these concerns and reassert its commitment to advancing AI in a responsible and open manner.

Source: AFP