Sources: UK-based Google DeepMind staff urge leaders to drop non-disparagement clauses barring public warnings about AI risks even after they leave the company

If you're working on AI, the ability to warn the public about risks without legal penalty is now a core employment condition.

Employees at Google DeepMind in the UK have reportedly urged company leadership to remove non-disparagement clauses from their employment contracts. These clauses allegedly prevent staff from publicly warning about the risks associated with artificial intelligence, even after their employment with the company has ended. Current workers expressed concerns that these contractual obligations are deterring individuals from speaking out about potential dangers and ethical considerations surrounding AI development. The internal push suggests a growing tension between the company's drive for AI innovation and employee-led calls for greater transparency and public accountability regarding AI safety.

Based on limited preview — the full article may require a subscription

The internal pressure from Google DeepMind staff highlights a critical debate within the AI industry regarding the balance between proprietary innovation and public safety. By seeking to remove these clauses, employees signal a desire for greater ethical accountability and a willingness to engage in public discourse about AI's societal impact. This move could set a precedent for how AI companies manage employee speech and the inherent risks of their technologies.

This situation underscores a potential shift in employee activism within the tech sector, where employees are increasingly prioritizing ethical considerations over traditional employment terms. If DeepMind concedes to these demands, it may compel other AI research labs and tech giants to re-evaluate similar clauses, potentially leading to a more open environment for discussing AI safety concerns across the industry.