OpenAI plans to train models on Daybreak Blue cybersecurity session data that was previously off-limits - the-decoder.com
Starting November 29, OpenAI will begin using Daybreak Blue cybersecurity session data to train its models, a change confirmed in an email to Trusted Access program members.
Previously, this data was excluded from training, even when users had the "improve the model for everyone" setting turned off. The policy change has sparked immediate backlash from researchers concerned about intellectual property protection.
Researcher Concerns
The change has researchers worried about a repeat of the Buckmaster-Alpöge case, where mathematicians uploaded Navier-Stokes equation drafts to OpenAI's Codex. The company later presented its own proof but denied using their inputs for training.
But the suspicion alone was enough to show researchers what could go wrong.
Safety Exceptions
Following the backlash, OpenAI issued a follow-up email promising a separate opt-out option. The company clarified that interactions flagged during safety reviews can still be used for safety purposes.
Related
