As organizations rapidly adopt artificial intelligence across their operations, customers, partners and other stakeholders want to know whether those systems are secure, reliable and properly governed. The AICPA is responding on two fronts. It has proposed significant updates to the SOC 2 Trust Services Criteria (TSC), and it has introduced a new assurance offering built specifically for AI agents.
What’s Changing in the Trust Services Criteria
In October 2026, the AICPA issued an exposure draft of revisions to the TSC, the criteria used to evaluate controls in a SOC 2 examination. If approved, the updated criteria will apply to SOC 2 examinations beginning on or after December 31, 2027, and early adoption is permitted. Public comments are due January 15, 2027.
The key changes include:
Restructured criteria: The common criteria are no longer organized around the 17 COSO principles. They are consolidated into eight series, and vendor management becomes a dedicated series. Service organizations will need to remap their existing controls to the new structure.
Updated points of focus: Points of focus have been revised throughout to reflect current risks and practices.
AI and emerging technology: The most notable update is that AI and other emerging technologies are built into the points of focus. The criteria now address AI wherever it affects the security, availability, processing integrity, confidentiality or privacy of a system.
AI Risk Is a Today Issue
The AI-specific points of focus won’t formally apply until examinations beginning on or after December 31, 2027, but the underlying risks already exist for any organization using AI within its system. The current TSC are risk-based and technology-neutral. If AI is part of how you deliver your services, its risks and the controls that address them are relevant to nearly all of the current criteria, including risk assessment, logical access, change management, system operations and vendor management. Examples include:
Unauthorized or unmonitored use of AI tools
Exposure of confidential data through AI services
Reliance on AI-generated outputs in processing
Service organization management is responsible for identifying these risks and designing controls to address them now. Waiting until the new criteria take effect leaves a gap in your control environment today.
SOC 2 + AIUC-1: A New Path to AI Assurance
Separately, the AICPA and the Artificial Intelligence Underwriting Company (AIUC) announced a joint offering that pairs a SOC 2 examination with AIUC-1 certification. Both are delivered through a single CPA-led engagement and result in one report.
AIUC-1 is a standard built specifically for AI agents. It was developed by a consortium of more than 250 CISOs and security leaders and is updated quarterly to keep pace with emerging AI risks. Its requirements span six principles:
Data and privacy
Security
Safety
Reliability
Accountability
Society
The standard maps to frameworks including MITRE ATLAS, ISO 42001 and the EU AI Act.
SOC 2 evaluates the controls around the systems that support AI. AIUC-1 addresses how the AI agent itself behaves. The certification requires technical testing of the agent at least quarterly. That testing includes adversarial attempts such as jailbreaks, prompt injection and data leakage.
The offering is designed for organizations deploying externally facing or high-impact AI agents with access to data, tools, systems or decision-making authority. It is currently in a pilot period, with a full launch expected in early 2027.
Preparing for What’s Next
Organizations with a SOC 2 report should take the following steps now:
Inventory AI usage: Identify where and how AI is used within your system boundary.
Assess current risks: Confirm AI-related risks are captured in your risk assessment and that controls are in place to address them.
Plan for the transition: Perform a gap assessment against the revised criteria ahead of your first report period under the new framework.
Organizations building or deploying AI agents should also consider whether a SOC 2 + AIUC-1 report fits their customers’ assurance needs.
Connect with Our Technology, Risk, and Compliance Professionals
Contact Our Technology, Risk & Compliance Professionals
Complete this form to discuss the updated Trust Services Criteria or SOC 2 + AIUC-1.
By submitting this form, you agree to be contacted by UHY.