Researchers Grow Weary of OpenAI's Closed-Door Approach to Breakthroughs
Expert 'Shepherds' to Decide How Breakthroughs Are Shared
Expert "shepherds" will decide how best to share the latest breakthroughs in computer science and physics, according to reporting by Robert Hart.
The Growing Frustration
Researchers are increasingly frustrated with OpenAI's practice of withholding its scientific breakthroughs from the broader research community. As the company has evolved from a nonprofit research lab into one of the world's most valuable technology firms, its findings in computer science and physics have become progressively harder to access.
Gatekeeping Under Scrutiny
The tension reflects a wider debate in the AI field heading into 2026: how much of the cutting-edge research conducted by leading labs should be shared openly with the scientific community. While OpenAI's charter commits it to broadly benefiting humanity, critics argue that commercial pressures have pushed the company toward secrecy that slows academic progress and limits independent verification.
'Shepherds' as Intermediaries
Under the reported arrangement, expert "shepherds" would serve as intermediaries, determining how best to disseminate the latest breakthroughs in computer science and physics. The approach aims to balance openness with safety and commercial considerations, though it remains to be seen whether researchers will view such gatekeeping as a meaningful solution.
What's at Stake
The outcome matters well beyond OpenAI. As AI systems grow more capable and research budgets balloon, access to foundational findings shapes who can build on them—and who cannot. With global AI competition intensifying and open-weight models from rivals gaining ground in 2026, the company's stance on sharing its work will remain a flashpoint between commercial strategy and the scientific community's expectations of transparency.
via The Verge AI