A group of current and former employees at two of the world's most powerful AI companies — OpenAI and Google DeepMind — went public on Tuesday with an open letter warning that the industry lacks adequate safety checks and that workers are effectively silenced from speaking out about risks.
The letter, titled "A Right to Warn about Advanced Artificial Intelligence," was signed by eleven current and former OpenAI employees and two from Google DeepMind. It was also endorsed by prominent AI researchers including Professors Yoshua Bengio and Stuart Russell, and Dr. Geoffrey Hinton.
What the Open Letter Says
The signatories argue that AI companies hold substantial non-public information about what their systems can and cannot do — including how well they protect data, what harms they could cause, and how significant those risks are. Yet companies currently face only weak obligations to share this information with governments, and no obligations at all to share it with civil society.
"We do not believe that they can all be expected to voluntarily share it," the letter states.
The authors warn of serious potential harms from advanced AI, including the entrenchment of existing inequalities, manipulation and misinformation, and in extreme scenarios, loss of control over autonomous AI systems.
The Whistleblower Problem
A central concern raised by the OpenAI and DeepMind employees AI safety letter is the near-total absence of whistleblower protections in the sector. Standard legal protections for whistleblowers focus on illegal activity — but many of the risks the signatories are worried about are not yet regulated.
"As long as there is no powerful government oversight over these corporations, current and former employees are among the few individuals who could be held personally responsible by the public," the letter states. "However, corporations need to keep us in the dark through strict non-disclosure policies, which in essence prevents us from expressing our concerns to anyone outside these organizations."
The letter lays out four core principles it asks AI companies to adopt:
- End non-disparagement clauses that bar employees from publicly discussing AI-related risks. - Create an anonymous internal process for workers to raise concerns directly with the board of directors. - Allow employees to report concerns to regulators and recognised AI safety watchdog groups without fear of retaliation. - Foster a culture of open criticism and commit to not retaliating against workers who go public after internal channels have failed.
OpenAI Responds; Google DeepMind Stays Silent
OpenAI pushed back, defending its track record and pointing to an existing tip-line where employees can report concerns. "We are proud of our work and will continue to approach the issue with science," an OpenAI spokesperson said. The company added that it would not launch new technology without creating adequate precautions first.
Google DeepMind did not issue a comment in response to the letter.
Context: A Wave of OpenAI Departures
The open letter arrived in the wake of several high-profile resignations from OpenAI and fresh scrutiny over the company's non-disclosure agreements. The departures and controversy had already drawn attention to questions of safety culture and internal accountability at the firm — making the letter's timing significant and its demands politically charged.
The signatories argue that until robust government oversight exists, the burden of holding AI companies accountable falls on the people who work inside them — and that those people need legal protection to do so.
