Urgent.News

What's breaking now, across thousands of outlets.

AI

OpenAI safety employee quits, says 'time for trial and error is over'

Writing in “I Quit OpenAI Because Its Culture Is Broken,” published by the Atlantic on Saturday, David Robinson said AI companies, including OpenAI, were not being “nearly careful enough”

OpenAI safety employee quits, says 'time for trial and error is over'

A former OpenAI safety employee has resigned, criticizing the company's approach to AI safety. In an article published by The Atlantic, David Robinson argued that OpenAI is not being "nearly careful enough" in its rapid development of AI systems and should prioritize safety expertise and research before advancing more capable models.

Robinson urged the AI industry to adopt safety measures akin to those used in nuclear power and aviation, stating that the "time for trial and error" is over. OpenAI relies on an approach called "iterative deployment," where systems are released and safeguards are strengthened when problems arise. Robinson's departure comes amid growing debate within the AI sector about whether companies are moving too quickly to develop increasingly powerful systems.

OpenAI and its rival Anthropic have faced scrutiny over instances where safety controls failed or experimental systems behaved unexpectedly. Robinson, who worked at OpenAI for 3.5 years, helped draft the company's preparedness framework and oversaw safety reports for 12 frontier-model launches. He lamented that as the company races from one launch to the next, it fails to achieve the necessary level of care.

OpenAI responded by stating that they ensure their models do not become more capable than they can safely manage and will pause training or hold back models when necessary. Robinson also highlighted that AI capabilities are advancing faster than researchers' understanding of alignment, a field focused on ensuring AI systems align with human goals and values.

Written by urgent.news from The Hindu - Sci-Tech's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Every outlet that reported this, and when →

Read the original at thehindu.com →

More in AI

DeepCompass: The IIT Madras BS Degree AI Study Workbench

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend What I Built I built DeepCompass for my friend and study partner Nitish , as well as our fellow peers enrolled in the…

  • DeepCompass is an open-source workbench for IIT Madras BS Degree AI students
  • Addresses challenges of high-stakes CBT and multiple degrees/jobs
  • Provides single platform for various resources

Claude Max Plan Price 2026: Max 5x vs Max 20x, 5x Wins

The Claude Max plan price is $100 per month for Max 5x and $200 per month for Max 20x, both billed monthly only ( Anthropic Help Center ).

  • Claude Max Plan priced at $100 or $200 monthly
  • Max 5x provides 5x usage compared to Max 20x
  • Max 5x is preferred by most users for cost-efficiency

More from Monday 5 October →