OpenAI’s NDAs Allegedly Restrict Employee Whistleblowing

Whistleblowers have approached the Securities and Exchange Commission (SEC), accusing OpenAI of using restrictive confidentiality agreements that impede employees from reporting safety concerns. The Washington Post obtained the exclusive complaint, which argues that OpenAI’s employment, severance, and nondisclosure agreements (NDAs) threaten penalties for workers who disclose issues to federal regulators. Confidentiality Agreements Under Fire The complaint asserts that OpenAI’s contracts force employees to give up their rights to whistleblower awards and require company approval before contacting federal authorities. The agreements purportedly lack exemptions for reporting securities violations to the SEC, potentially violating federal protections for whistleblowers. In a recent letter to the SEC, the whistleblowers expressed that these contracts deter workers from engaging with federal regulators. A whistleblower emphasized that AI companies can’t ensure safety if they shield themselves from criticism and oversight. OpenAI’s Official Stance OpenAI spokesperson Hannah Wong told The Post that the company’s whistleblower policy supports employees’ rights to make protected disclosures. Wong also noted recent changes, including the removal of nondisparagement clauses from departure agreements. The complaint surfaces amid worries that OpenAI, initially a nonprofit committed to benefiting humanity, may now prioritize financial gain over safety. The Washington Post noted that OpenAI launched its latest AI model swiftly despite internal concerns about the need for more rigorous security checks. Lindsey Held from OpenAI denied compromising safety but acknowledged the pressure faced by the team during the release. SEC Action Demand In correspondence to SEC Chairman Gary Gensler and Congress, the whistleblowers urged the SEC to investigate OpenAI’s allegedly unlawful agreements. According to whistleblowers’ attorney Stephen Kohn, the SEC has responded to the complaint, though it is unclear if a formal investigation has begun. The letter calls for the SEC to mandate that OpenAI disclose all agreements with nondisclosure clauses and notify employees of their rights to report violations confidentially. Additionally, it recommends imposing fines for each noncompliant agreement and requires OpenAI to take steps to counteract the chilling effect of prior contracts. Recent Employee Pushback Last month, reports suggested employees are concerned about what OpenAI is doing with its substantial equity. In accordance with the company’s publicly announced policy, annual tender offers are made to employees, granting them the opportunity to sell a predetermined portion of their equity holdings. However, a recent review of internal documents revealed a stipulation that excludes former employees currently employed by competitor firms from participating in these sales events. This restriction has caused concern among existing shareholders, who now express anxieties regarding their ability to divest themselves of their equity holdings within the established framework. Also in June, a group of former OpenAI employees and existing Google DeepMind employees wrote an open letter showing concern over safety measures in AI companies. The group of expressed concern that OpenAI’s focus on growth and profitability may be neglecting critical issues of safety and transparency, especially in the context of developing artificial general intelligence (AGI). Elsewhere, Leopold Aschenbrenner, an ex-safety researcher at OpenAI, publicly claimed his termination was driven by his efforts to spotlight AI security flaws within the organization.