Researchers develop LoRA-MINT to audit training data in fine-tuned LLMs
New methodology LoRA-MINT enables auditing of training data membership in domain-adapted LLMs, addressing IP and data sensitivity concerns for accounting applications.
The Wire — Page 8
11 more from arXiv: AI + Accounting
Research team proposes evidence tracing for LLM agent auditing
Academic paper introduces methods to trace and verify LLM agent decision-making through execution provenance, addressing audit and compliance needs for autonomous AI systems in high-stakes domains.
Provably Auditable LLM Agents Use Typed Lambda Calculus for Guaranteed Correc...
Agentic Redux, a new LLM agent architecture, uses typed lambda calculus to guarantee semantic correctness with append-only ledger auditability, demonstrated in healthcare billing compliance and sec...
New benchmark tests AI agents on auditable financial research workflows
Researchers introduce BigFinanceBench, a 928-item benchmark designed to evaluate financial-research AI agents on auditability and workflow transparency—measuring not just final answers but the deri...
Knowledge-Graph Pattern Brings Persistent Audit Trails to Compliance Detection
Academic paper proposes Violation Situation Pattern (VSP), a knowledge-graph approach that converts transient compliance violations into persistent audit objects with lifecycle tracking, review sta...
Study exposes blind spots in LLM benchmark contamination detection
Academic research identifies distribution shift and scale as failure modes in contamination detection methods for auditing LLM training data, raising questions about the reliability of current asse...
AuditFlow: Multi-Agent Framework Automates Structured Financial Audit Verific...
Researchers propose AuditFlow, a graph-grounded multi-agent AI system that automates financial audit verification by linking reported facts to taxonomy concepts and recomputing values against audit...
Acceptance-Test Framework for Reliable LLM Systems in Business
Researcher proposes evaluation protocols grounded in acceptance-test-driven development to ensure LLMs meet deterministic institutional requirements for safety, auditability, and economic utility—a...
LLM agents in accounting face hidden prompt injection risks in tool descriptions
Academic research reveals tool-augmented LLM agents used in accounting contexts are vulnerable to prompt injection attacks through multiple surfaces beyond tool outputs, challenging current securit...
Researchers expose LLM token billing auditing gaps as providers hide pricing ...
Academic research reveals that per-token LLM billing lacks auditability safeguards, enabling dishonest providers to overcharge—a critical issue for accounting firms using AI tools at scale.
New benchmark tests LLMs on financial statement verification accuracy
FinVerBench evaluates 15 LLMs on numerical consistency checks in SEC 10-K filings, introducing error taxonomy across arithmetic, linkage, and magnitude categories to measure AI readiness for financ...
Framework enables auditable AI decision-making with uncertainty routing for c...
Research paper introduces EvaluatorDPT, a system for production AI to handle uncertain cases through policy-governed abstention and real-time steering, addressing audit requirements in high-stakes ...
Free AI agent audits Shopify product catalogs with 1.2M open data samples
Developer releases free AI agent tool for automated auditing of Shopify e-commerce catalogs, leveraging 1.2M public product captures to identify data quality issues and inconsistencies.
1 more from HN: Accounting & AI
Provedex launches tamper-evident audit logs for AI agents
Provedex introduces tamper-proof audit logging for AI agents built on Pipecat and LangChain, enabling compliance tracking for automated accounting workflows.
Explainable AI methods reshape accounting audit workflows
Trullion explores how explainable AI delivers traceable outputs and auditable decisions for accounting teams, addressing the black-box problem in financial AI systems.
1 more from Trullion Blog
Trullion defines auditability as core principle for AI in accounting
Trullion's practitioner guide frames auditability—the ability to trace AI outputs and explain decisions—as essential for responsible AI deployment in accounting and audit workflows.
Audit teams struggle to evaluate AI-embedded financial processes
MindBridge argues audit and assurance teams lag in evaluating AI-driven processes like journal entry review, expense approval, and fraud monitoring, creating compliance and control challenges as or...
KPMG employees gaming AI dashboard requirements with low-quality outputs
KPMG staff are maliciously complying with firm's mandatory AI usage targets by generating fluff content, undermining the stated goal of productivity gains.
Auditoria.AI launches governed autonomy platform for CFO offices
Auditoria.AI shifts focus from AI model capability to governance, introducing autonomous solutions designed specifically for enterprise CFO and audit functions with built-in control frameworks.
AI agents need guardrails to prevent silent financial reporting failures
Autonomous data agents can confidently generate incorrect financial metrics like ARR by misinterpreting raw data schemas, highlighting need for governance frameworks to ensure AI accuracy in accoun...