Researchers question whether OpenAI can be trusted with unpublished math shared through its AI systems

The post links to several social media discussions raising concerns about whether mathematicians should share unpublished work with OpenAI tools. The central issue is whether sensitive research inputs could be retained, used for training, or otherwise exposed before publication.

The debate highlights a broader trust gap between frontier AI labs and researchers handling unpublished work. As AI tools become more useful for advanced math, adoption may depend less on capability and more on enforceable privacy, data-use, and attribution guarantees.