Leading AI systems now complete medium-length, well-defined accounting tasks more quickly and accurately than entry-level human professionals, including the top performers in a recent study. The research indicates that frontier models have effectively bridged the gap in precision and speed when handling specific financial exercises.

The results mark a reversal from 18 months ago, when the most advanced AI models fell short of an average human accountant's score of approximately 37%. The study's authors nearly withheld the provocative findings for fear of misinterpretation but decided to publish them to help institutions prepare for the pace of AI advancement.