Trust standards for the AI era are taking shape. In the future, leading companies will need not only AI capabilities but also to demonstrate why the AI acts as it does and who is accountable for it.Author and source: Sumsub
AI is entering the era of autonomous execution.
From intelligent customer service and automated operations to trade analysis, risk assessment, and compliance review, AI agents are gradually evolving from auxiliary tools into independent digital entities capable of completing tasks on their own.
But as AI autonomy increases, a new issue is emerging: when AI makes decisions on behalf of a company, can the company demonstrate why it did so, who authorized it, and on what basis?
In August 2026, Sumsub, a global identity verification and digital trust platform, partnered with the Singapore FinTech Association to release the report "State of Digital Trust in Asia-Pacific: AI Governance Benchmark."
The report surveyed nine markets in the Asia-Pacific region and 720 professionals from risk, compliance, technology, and operations fields, providing an in-depth analysis of the new governance challenges enterprises face in the age of AI agents.
The report introduces a key concept: "accountability asymmetry," meaning that companies are held responsible for the outcomes generated by AI, yet may not be able to demonstrate how AI decisions were reached.
AI applications are accelerating, but governance capabilities have not kept pace.
The report shows that Chinese enterprises are rapidly accelerating their adoption of AI:
- Approximately 97% of Chinese enterprises are using or piloting multi-stage AI systems.
- However, only 29% of companies have sufficient confidence in tracking and explaining AI decision-making processes.
This means that many businesses have begun relying on AI to perform complex tasks, but they still lack complete verification capabilities regarding the decision pathways, data foundations, authority sources, and operational logs underlying AI behavior.
In the past, companies had to prove: “Why did this employee make this decision?” In the future, companies will need to answer: “Why did this AI agent make this decision?”
The core of AI governance: making autonomous behavior verifiable
Sumsub evaluates APAC enterprises' AI governance capabilities across three dimensions:
1. Autonomy:
The extent to which AI can complete tasks independently.
With the advancement of agentic AI, AI is moving beyond simple content generation to automatically executing workflows, invoking external tools, making complex decisions, and interacting with other systems.
2. Responsibility:
After AI generates results, is there a clearly defined responsible party?
The report shows that 75% of Chinese enterprises have assigned responsibility for AI-related outcomes to specific individuals or teams: 35% to individuals and 40% to teams. This indicates that companies are establishing AI accountability frameworks.
3. Traceability:
Can enterprises record and reconstruct AI decision-making processes?
This is currently the biggest challenge. Only a few companies have the capability to maintain complete AI decision-making records. If AI agents are involved in financial transaction assessments, user risk ratings, identity verification, and fraud detection, yet cannot explain the basis for their decisions, companies will face new trust risks.
From Digital Identity to AI Identity: The Next Generation of Trust Infrastructure Is Taking Shape
As AI agents become important participants in the digital economy, a new question arises: Do AI agents also need identities?
In the traditional internet era, identity systems answered: "Who is this person?" In the AI era, we must further answer: "Who is this AI? What is it authorized to do? What has it done?"
The report states that future enterprises will need to establish:
- AI Agent Identity Management
- Access control mechanism
- Behavior logging system
- Decision audit capability
Ensure that every AI operation is traceable to a verified AI entity and the responsible individual behind it.
In the era of Web3 and AI convergence, trust becomes a critical infrastructure.
This trend is particularly evident in the Web3, digital assets, and fintech industries.
Decentralized applications, smart contracts, digital asset trading, and AI agent services all rely on a core capability: trusted identity.
The future digital economy requires not only smarter AI and more automated systems, but also more transparent authorization mechanisms, more reliable identity systems, and enhanced auditing capabilities.
Enterprises can only achieve true large-scale adoption when AI behavior is verifiable.
From “being able to use AI” to “proving AI is trustworthy”
The report shows that 99% of companies are willing to adopt methods that can verify each AI operation.
This indicates that the market is forming a new consensus: the next phase of AI competition is not just about model capabilities, but about who can establish a trustworthy AI governance system.
As Penny Chai, Vice President of Sumsub for Asia-Pacific, stated:
Regulatory rules clearly define responsibility boundaries, while digital trust infrastructure such as identity verification, decision tracking, and audit logs enables AI accountability to be effectively implemented in real-world applications.
Trust standards for the AI era are taking shape. In the future, leading companies will need not only AI capabilities but also to demonstrate why the AI acts as it does and who is accountable for it.
Download the full report
Want to learn more about AI governance trends in Asia-Pacific, the current state of enterprise AI agent adoption, and how to build verifiable digital trust? Click the original link to access the "Sumsub Digital Trust Status in Asia-Pacific: AI Governance Benchmark Report."