Former DeepMind PR Staffer Says Lab Banned Public Discussion of AI Extinction Risk
A former public relations staffer at DeepMind has revealed that the company once prohibited employees from publicly discussing the risk of AI causing human extinction. The policy, according to the ex-employee, was designed to avoid negative media coverage and public scrutiny.
The individual, who worked in communications at the Google-owned lab, claimed that internal guidelines explicitly forbade mentioning “extinction risk” or similar catastrophic scenarios in external communications. This included blog posts, interviews, and even private conversations with journalists.
The Ban Came From Leadership
“We were told point-blank not to talk about extinction. It was seen as bad for our brand and would scare away partners and investors.”
The former staffer said the directive came from senior leadership, not from a single department. The goal was to maintain a positive image of DeepMind as a responsible, safety-focused organization without drawing attention to worst-case outcomes.
DeepMind has not commented directly on this specific allegation. However, the company has repeatedly stated that it takes AI safety seriously and has published research on alignment and governance.
Why the Ban Matters for AI Governance
The revelation highlights a tension between public relations and honest risk communication in the AI industry. If labs suppress discussion of existential threats, experts warn that policymakers and the public may remain underinformed.
- Censorship of risk discourse can slow down regulatory action. Without open debate, lawmakers lack the pressure needed to pass strong safety laws.
- Internal fear of backlash may prevent researchers from publishing safety-critical findings. Several current and former employees have complained about limited freedom to discuss worst-case scenarios.
- Reputation management often overrides transparency. In a high-stakes field like AI, the urge to control the narrative can conflict with the need for democratic accountability.
DeepMind’s Broader Safety Record
DeepMind has long positioned itself as a pioneer in AI safety. It spun off an independent safety research group, issued public pledges, and collaborated with regulators. Yet this PM staffer’s account suggests that behind the scenes, the lab avoided publicizing the most alarming potential outcomes.
Other former employees have made similar claims. For example, some have said that DeepMind managers discouraged papers on “AI doom” and asked researchers to focus on more optimistic scenarios.
What the Ban Meant for Staff
Those who worked at the lab during the period described a chilling effect. Researchers hesitated to raise existential-risk concerns in internal meetings, knowing that public statements would be vetted or silenced.
- Junior researchers self-censored to protect their careers.
- PR teams actively rewrote or blocked statements that mentioned extinction.
- External partners were given curated messaging that omitted worst-case scenarios.
The Bigger Picture: Industry-Wide Secrecy
DeepMind is not the only lab facing such criticism. OpenAI, Anthropic, and others have been accused of downplaying existential risk to maintain investor confidence. The former PR staffer’s testimony adds weight to calls for mandatory transparency requirements in advanced AI development.
As AI capabilities accelerate, the gap between what companies say publicly and what they know privately may become a central policy issue.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.