David Robinson, a former safety reports writer at OpenAI, has resigned and is now sounding the alarm in an editorial in The Atlantic. He claims the culture in the industry is ‘broken’, suggesting that the relentless focus on progress over caution could be deeply problematic.

Robinson argues that Silicon Valley's 'extreme confidence' and 'unimpeded optimism' are misguided, and that AI companies need to adopt a more cautious and humble approach, akin to nuclear power plants or busy airports, with layers of redundancy and careful planning to prevent disasters.

His concerns are echoed by a growing number of ex-researchers and safety workers, including Jacob Coxon from Anthropic and Robert O’Callahan from Google DeepMind. They are all pushing for a more cautious approach, warning that the risks of AI could be far more significant than previously thought.

The departure of these experts from major AI firms highlights a growing unease within the industry. While some might dismiss these warnings as sour grapes, the sheer number of departures suggests a systemic issue that cannot be ignored.

As AI becomes increasingly integral to our lives, the need for robust safeguards is becoming more pressing. Will we learn from these warnings or continue down the path of potential disaster?