OpenAI Under Investigation: NDAs Can Silence Whistleblowers

Former employees of OpenAI filed a complaint with the United States Securities and Exchange Commission (SEC) requesting an investigation into the company's alleged overly restrictive non-disclosure agreements (NDAs) artificial intelligence (AI), according to a letter obtained by Reuters.

“Considering the well-documented potential risks posed by irresponsible implementation of AI, we urge Commissioners to immediately approve an investigation into the company’s previous NDAs. OpenAI and review current efforts apparently undertaken by the company to ensure full compliance with SEC rules,” says the letter, provided to Reuters by the SEC’s office. Senator Chuck Grassley.

According to the complaint, the OpenAI would have forced employees to sign agreements that forced them to give up their federal rights to whistleblower compensation. The whistleblowers requested that the SEC fine the OpenAI for each improper agreement, to the extent the agency deems appropriate.

“Artificial intelligence is rapidly and drastically altering the technology landscape as we know it,” said Senator Grassley, whose office said it received the whistleblowers’ letter. He added that “the policies and practices of OpenAI appear to have a chilling effect on the right of whistleblowers to speak out and receive appropriate compensation for their protected disclosures.”

The news was initially reported by the Washington Post.

The complainants allege that the OpenAI issued overly restrictive employment, termination and confidentiality agreements, which could have led to penalties against employees who raised concerns about the OpenAI to federal authorities.

The letter also states that the OpenAI required employees to obtain prior company consent to disclose information to federal regulators, and that did not create exceptions in non-defamation clauses for reporting securities violations to the SEC.

Furthermore, the letter asks the SEC to require the OpenAI present all contracts that contain a nondisclosure agreement, including employment, termination, and investor contracts, for inspection.

As AI models become more powerful, chatbots in OpenAI with generative AI capabilities such as humanized conversations and text-based image creation have raised security concerns.

In May, the OpenAI formed a Security Committee that will be led by board members, including the CEO Sam Altman, for training your next artificial intelligence model.

Read also

OpenAI develops new reasoning technology called “Strawberry”