The AICPA and Artificial Intelligence Underwriting Company developed the first assurance certification for AI systems and agents, currently in pilot phase ahead of full launch in early 2027.
New framework uses AI agents to discover and version data quality rules with audit trails, addressing traceability gaps in high-stakes financial data pipelines subject to regulatory oversight.
Academic paper introduces framework for auditors to verify which LLM model actually processed requests in hosted services, addressing opacity in AI model deployment claims without requiring runtime...
Research reveals LLM agents with shared long-term memory expose other users' credentials 84.7% of the time, proposing audit framework to distinguish between exposure, misuse, and account compromise.
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...
TamperAudit introduces tamper-evident audit logging for AI agents to comply with EU AI Act Article 12 requirements, addressing regulatory demand for transparent, immutable records of agent decisions.
AgentBlackBox launches a policy gate and audit logging system for AI agents, addressing compliance and accountability needs for automated systems in regulated industries including accounting.
Makro's GI Agenda Tributária is an AI agent designed to help accountants and tax professionals track tax obligations and deadlines automatically, reducing manual compliance work.
Xero outlines how agentic AI can streamline fragmented accounting processes—spend controls, fraud checks, approvals, reconciliation—into a single cash flow system for small businesses managing payr...
New arXiv paper introduces PROBE, a method for evaluating whether tool-using LLM agents comply with organizational policies on pricing, privacy, and safety—addressing a gap between capability bench...
Academic paper introduces Functional Agent Operating Procedures (FAOPs) with formal semantics to make multi-step AI agent workflows traceable, auditable, and compliant—addressing mandatory requirem...
Researchers propose a relational learning approach to detect money-mule accounts by analyzing network patterns and transaction shapes rather than individual transfers, addressing a gap in productio...
Researchers identify critical security flaws in AI agent authorization workflows, demonstrating 12 attack vectors (decoy recipients, amount cancellation, label spoofing) that can bypass approval me...
AICPA issues guidance clarifying how audit standards apply when entities use AI in financial statements and auditors deploy AI during audits.
Optro unveiled a GRC Intelligence Platform featuring domain-trained agents designed for audit, compliance, third-party risk, and AI governance with human-led control.
AICPA releases practice aid addressing two critical audit scenarios: when clients use AI in financial reporting/controls and when auditors deploy AI technology in audit procedures.
CPA Practice Advisor examines critical written policies firms need before using AI tools with client tax data, focusing on data location and control governance.
Tenuo, a task-scoped authorization framework for AI agents, defended a fork of OWASP's financial chatbot against 50 promptfoo attacks by adding deterministic guardrails that prevent poisoned instru...
MindBridge outlines a four-layer accountability stack for AI agents handling financial workflows, addressing how automation can hide risks in invoice approvals, journal entries, and transaction pat...
Auditoria.AI reports 66.5% of finance orgs increasing AI investment, with new research on deploying AI agents across AP/AR processes and finance controls.