According to reports, OpenAI published more than 700 pre-prints on GitHub on October 6 describing the significant mathematical results achieved by its internal AI model. Less than 24 hours later, the company has withdrawn three of these manuscripts due to errors. OpenAI released 722 preprints, claiming to have made progress on 372 math problems in geometry, computer science, algebra, and other disciplines. Hundreds of new mathematical discoveries generated by artificial intelligence have advanced the cutting edge of advanced mathematics in a single day. However, on October 7, OpenAI announced the withdrawal of three manuscripts due to a symbol error. According to the announcement, the error invalidated one of the arguments in one manuscript and invalidated the constructs used in the two reliant papers. Each retracted paper was accompanied by a description explaining the vulnerability, and stated that OpenAI had “revised another 14 manuscripts, including proof fixes, statement corrections, clearer assumptions and dependencies, and a correction to an outdated citation.”

Disclaimer:Webull uses external vendor Google Translation Service for news translations where we endeavour to ensure these are correct, however, we recommend that you please double-check this information accordingly. Webull is not responsible for translation errors or issues.