OpenAI fires 3 researchers; UC Merced and Merced College explain AI rules

Key Takeaways

OpenAI said Oct. 9 it fired three researchers for mishandling sensitive information.
UC Merced formed an AI Advisory Council in 2026 and lists approved AI tools with data levels.
Merced College offers a 6–8 hour self-paced AI course on responsible AI use for faculty and staff.
Local institutions caution against putting P3 or P4 data into unsanctioned AI tools.

It was Friday, Oct. 9. OpenAI said it had fired three safety researchers after an internal investigation into how sensitive company information was handled. The details are contested. For Merced, the question lands closer to home than Silicon Valley drama: what do students, staff, and local employers do to keep their own data tight?

What happened Friday

OpenAI announced the dismissals Friday morning, saying the employees violated policies on handling sensitive information. The fired researchers disputed the account in public posts over the weekend. None of that has much to do with almond prices or the bus schedule in Atwater, but it does set the tone for how the biggest AI companies talk about safety and accountability.

Why it matters here

UC Merced has been building its own playbook this year. The university created an AI Advisory Council in February 2026 and now keeps a public list of approved tools, from Microsoft Copilot to campus-built assistants, with clear labels for what institutional data each tool can touch. The guidance is blunt about one thing: if a tool isn’t covered by a campus or UC contract, be careful with it, especially anything that looks like student records or health data. A vending machine hummed in the OIT lobby on Monday afternoon.

Merced College has taken a practical tack. The district runs a 6–8 hour self‑paced course for faculty and staff on how to use generative AI in class and at work, with a focus on data handling and when to say no. The college’s materials point staff to tools available through the California Community Colleges’ partnerships, and they draw the same bright line around sensitive information.

What local rules say

UC Merced’s AI site ties approved tools to the university’s data protection levels, P1 through P4, so a department can see at a glance where a meeting transcript or a draft grade report fits. The campus also notes that ChatGPT EDU licenses are in the pipeline systemwide, which, once live, would put stronger contractual safeguards around a popular tool. But the basics are clear.

One sentence reads: "Students should not use AI for coursework without explicit permission from their instructors."

Central Valley AI is produced by the CVAI Newsdesk team and developed by Kaweah Tech, a regional firm that builds, deploys, and integrates AI solutions for businesses across California's Central Valley.

Source

https://www.mercedsunstar.com/news/business/article317600234.html