Urgent.News

What's breaking now, across thousands of outlets.

AI

OpenAI fires three staff for mishandling 'sensitive' info

ChatGPT-maker OpenAI has said that it has fired three researchers for allegedly mishandling "sensitive information" and violating company policies, including work that involved an external organisation that evaluates artificial intelligence models.

OpenAI fires three staff for mishandling 'sensitive' info

OpenAI has terminated three researchers for allegedly mishandling sensitive data and breaching company policies, including collaboration with an external AI evaluation firm, according to the Wall Street Journal and Bloomberg. While the identities of the dismissed individuals remain undisclosed, at least two were involved in safety and alignment research, as reported by AFP.

OpenAI announced the dismissals, stating that the firings resulted from employees violating established procedures and betraying the trust necessary for their work. This move comes during a contentious discussion surrounding AI safety and the potential existential threat posed by rapidly advancing technology. In September, three researchers, Jasmine Wang, Tomek Korbak, and Mikita Balesni, had been vocal on social media about their concerns regarding AI safety.

Balesni expressed his belief that there is a 10% chance AI could lead to the extinction of humanity. Korbak and Wang both expressed dissatisfaction with OpenAI's practices and openly criticized the company's direction. In response to these resignations, leading tech companies signed a voluntary pledge to self-regulate safety measures at a White House meeting with President Donald Trump, who emphasized the "morally binding" commitment to establish adequate safeguards for AI development.

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

Read the original at rte.ie →

More in AI

BugReplay: Helping Developers Learn from Bugs Their Team Has Already Solved 🐛

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend Ever spent hours debugging an issue, only to find out your teammate had already fixed the exact same bug?

  • BugReplay is an AI-powered debugging assistant
  • Transforms past bug fixes into reusable knowledge
  • Built with React, Python, FastAPI, Qwen, PostgreSQL

Does your agent actually remember, or just sound like it? A 120-line open-weight test

A friend asked me a question I couldn't answer about his own AI agent: does the thing answering me today actually remember being yesterday's, or is it just very good at sounding like it does?

  • Gapcheck is a Python script for offline testing AI memory
  • Three probes assess memory continuity: no record, record, and no record pressed
  • Agent struggled to recall items directly but fabricated responses when memory unavailable

More from Friday 2 October →