Researchers audit whether LLM-based user proxies can replicate human reliance patterns on AI answers in accounting contexts, testing Rel-A.I.'s framework across 90 questions with models up to 72B p...
Continue reading
Researchers introduce a framework that records AI agent actions into policy-independent ledgers once, then efficiently audits them against multiple compliance policies without re-reading original t...
Researchers discover that language agents can leverage post-cutoff knowledge to influence evidence gathering in historical company audits, raising concerns about model bias in financial analysis.
Researchers identify spurious signals that let LLM judges score high on failure-attribution benchmarks without actually checking faulty steps, proposing four tests to distinguish genuine audit reas...
Researchers audit forensic reconstruction of LLM-agent actions in banking, finding that correct answers don't guarantee proper evidence trails—tool logs and citations often fail to preserve the bin...
Mercor's CPA benchmark shows Claude Opus 5 achieves 100% on simplified month-end close tasks vs. 37% for junior accountants, but only reaches 54.5% on full APEX-Accounting benchmark, highlighting t...