OpenAI employees forced regulators to sign illegal non-disclosure agreements to waive whistle-blowing rights
by
Anonymous OpenAI employees have written a letter to the U.S. Securities and Exchange Commission (SEC) asking it to investigate OpenAI for alleged misconduct, The Washington Post reported.
OpenAI illegally stopped staff from sharing dangers, whistleblowers say - The Washington Post
https://www.washingtonpost.com/technology/2024/07/13/openai-safety-risks-whistleblower-sec/
Whistleblowers accuse OpenAI of 'illegally restrictive' NDAs | TechCrunch
https://techcrunch.com/2024/07/13/whistleblowers-accuse-openai-of-illegally-restrictive-ndas/
According to a seven-page (PDF) whistleblower document obtained by The Washington Post, OpenAI forced its employees and investors to sign 'overly restrictive and unlawful non-disclosure agreements (NDAs) ' that prohibited them from whistleblowing or waived their rights to do so.
The specific illegal activities cited by OpenAI employees are as follows:
- The 'company defamation clause' does not include any exclusion from reporting to authorities.
- Require prior consent from the company to disclose confidential information to authorities.
- Confidentiality agreements that themselves contain breaches of regulatory law.
- Encouraging whistleblower reporting and forbidding whistleblower compensation.
The employees pointed out that such practices not only violate laws prohibiting the obstruction of reporting securities violations, but also lead to the inappropriate use of AI, which could lead to serious social harm, including false information, fraud, bias, and national security risks. They called on the SEC to promptly investigate and take corrective measures.
A copy of the letter was also sent to the U.S. Congress, in which Senator Chuck Grassley said, 'Congress' constitutional responsibility for national security is to monitor and mitigate threats posed by AI, and whistleblower support is essential to carrying out that responsibility. However, OpenAI's policies and practices risk undermining whistleblowers' right to speak out and to receive fair compensation for their reporting.'
The whistleblower's report comes amid concerns that OpenAI, which started as a non-profit organization, is prioritizing profits over safety in its technology development.
The Washington Post previously reported that 'OpenAI rushed to release GPT-4o in May 2024, despite the failure of its own protocols to prevent its AI from teaching users how to create biological weapons or allowing hackers to develop new cyber attacks.' In response, OpenAI said, 'We did not compromise our safety processes, but we know this release was stressful for our team.'
The letter is dated June 1, 2024, just one month after GPT-4o was released, but it is unclear at the time of writing whether the SEC has launched an investigation. The SEC declined to comment when contacted by media.
OpenAI has long been accused of coercive employee practices, for which CEO Sam Altman has apologized.
OpenAI CEO Altman apologizes for 'aggressive tactics' used against departing employees in internal documents - GIGAZINE
by steamXO
Related Posts:
in Note, Posted by log1l_ks