Whistleblowers denounce 'illegal' confidentiality agreements imposed by OpenAI on employees

Whistleblowers have accused OpenAI of imposing “illegal” restrictions on how employees communicate with government regulators, according to a letter obtained by OpenAI. Washington Post.

According to the letter to SEC Chairman Gary Gensler, OpenAI, known for its AI technology including ChatGPT, would use “illegal” confidentiality agreements.

A traditional confidentiality agreement generally prohibits an employee from disclosing company secrets to competitors and the public.

However, the agreements OpenAI will sign with its employees will prevent them from communicating with federal authorities regarding security concerns, according to the anonymous whistleblower.

The latter has therefore asked the US financial supervisory authorities to investigate these agreements, alleging “systemic” legal violations within the startup.

According to the letter, a formal complaint has also been filed with the Securities and Exchange Commission.

“The agreement prohibits and discourages employees and investors from communicating with the SEC regarding security breaches…and requires employees to notify the company when they communicate with government regulators,” the representative reportedly said in the letter.

Given the potential danger posed by AI, these whistleblowers stressed the importance of regulating this technology, as well as the role of employees and whistleblowers in national security.

“There is an urgent need to ensure that employees working on this technology understand that they can raise complaints or raise concerns with federal regulatory authorities or law enforcement,” they wrote.