Tenon AI Security Network Launched by Terminal 3 and Affinidi
The briefing, recorded by CNA and featuring Terminal 3 chief executive Gary Liu, highlighted the network’s initial focus on the financial services sector and described the technical safeguards designed to protect against cyber threats.
In a joint announcement made at the Singapore FinTech Festival on 19 November 2026, Singapore‑based digital technology firm Affinidi and Hong Kong‑based deep‑tech company Terminal 3 unveiled Tenon, a public network intended to verify and secure artificial‑intelligence agents that process sensitive data for organisations. The briefing, recorded by CNA and featuring Terminal 3 chief executive Gary Liu, highlighted the network’s initial focus on the financial services sector and described the technical safeguards designed to protect against cyber threats. The development of Tenon arrives at a moment when Japanese regulators, industry bodies and technology firms are intensifying scrutiny of AI security, making the partnership particularly relevant to Japan’s broader AI governance agenda.
Strategic Rationale Behind the Tenon Partnership
The collaboration between Affinidi and Terminal 3 reflects a strategic alignment of complementary capabilities. Affinidi, founded with investment from Singapore’s sovereign wealth fund Temasek, brings a portfolio of digital identity and credential‑verification tools that have been applied across Southeast Asian markets. Terminal 3, positioned in Hong Kong’s deep‑tech ecosystem, contributes expertise in building large‑scale, low‑latency networks for high‑frequency data exchange. By joining forces, the two firms aim to create a network that can certify AI agents in real‑time, thereby reducing the risk of data leakage or manipulation when AI systems interact with confidential information.
The choice of the financial services sector as Tenon’s launch arena underscores the perceived vulnerability of that industry to AI‑related threats. Financial institutions routinely deploy AI for fraud detection, credit scoring and algorithmic trading, all of which involve the handling of personally identifiable information and transaction data. A public verification network promises to add a layer of trust, ensuring that AI agents operating within a bank’s ecosystem have been authenticated and are operating within defined security parameters.
For Japan, whose banking and securities sectors are undergoing rapid digital transformation, the Tenon model offers a potential template for public‑private cooperation on AI security. The Ministry of Economy, Trade and Industry (METI) has recently emphasized the need for standards that can be applied across institutions, and the Tenon initiative could inform future Japanese policy discussions on interoperable AI verification frameworks.
Technical Architecture and Verification Process
According to the interview with Gary Liu, Tenon operates as a public network that issues cryptographic attestations to AI agents. When an AI model is deployed, it must register with the Tenon network, providing metadata about its training data, intended use case and security controls. The network then generates a digital certificate that can be queried by any participating organisation to confirm the agent’s authenticity and compliance status.
The verification mechanism relies on a combination of zero‑knowledge proofs and secure enclaves, allowing the network to confirm that an AI agent meets prescribed security standards without exposing proprietary model details. This approach addresses a common concern among Japanese firms that wish to protect intellectual property while still participating in a shared security infrastructure.
In addition to the attestation process, Tenon incorporates continuous monitoring. AI agents are required to periodically submit health checks, which include logs of data access patterns and anomaly detection signals. Should an agent exhibit behaviour that deviates from its baseline, the network can flag the anomaly to the owning organisation and, if necessary, revoke the agent’s certificate. This dynamic revocation capability is intended to keep pace with the rapid evolution of AI capabilities and the corresponding threat landscape.
Cyber‑Threat Safeguards Embedded in Tenon
Security safeguards form a core component of Tenon’s design. Liu emphasized that the network is built to resist a range of cyber‑threat vectors, including man‑in‑the‑middle attacks, credential theft and model poisoning. By using end‑to‑end encryption for all certificate exchanges, Tenon ensures that only authorised parties can verify an agent’s status.
Model poisoning, where adversaries manipulate training data to embed hidden behaviours, is a particular focus. Tenon requires participating organisations to submit provenance records for the data used to train AI agents. These records are stored in an immutable ledger, enabling auditors to trace the origin of training datasets and detect inconsistencies that could indicate tampering.
For Japanese financial institutions, which are subject to stringent cybersecurity regulations enforced by the Financial Services Agency (FSA), the ability to demonstrate compliance with such provenance requirements could become a competitive advantage. The network’s transparent audit trail aligns with Japan’s push for stronger cyber‑resilience in the financial sector, as outlined in recent FSA guidelines.
Implications for Japan’s AI Governance Landscape
The launch of Tenon arrives as Japan finalises its AI strategy for the 2027‑2032 period. The government has signalled a desire to balance innovation with robust risk management, and the Tenon model provides a concrete example of how verification infrastructure can be institutionalised. METI’s AI policy unit may consider adopting similar public‑network concepts to support domestic AI developers, particularly those engaged in cross‑border data processing.
Moreover, the public nature of Tenon could dovetail with Japan’s efforts to harmonise AI standards internationally. By participating in a network that includes firms from Singapore and Hong Kong, Japanese companies would gain exposure to regional best practices, potentially informing the development of domestic standards through bodies such as the Japan Electronics and Information Technology Industries Association (JEITA).
From a regulatory perspective, the network’s certificate‑based approach may simplify compliance reporting. Instead of each institution maintaining its own separate verification system, a shared public network could serve as a trusted third‑party source of evidence, reducing administrative burdens on Japanese firms while enhancing overall system security.
Challenges of Keeping Security Infrastructure Current
While Tenon offers a promising framework, Liu acknowledged the difficulty of maintaining security infrastructure that can keep up with rapid AI advances. As AI models become more complex and capable of autonomous decision‑making, the criteria for certification will need continual revision. This creates a demand for agile governance processes that can update verification standards without causing disruption to operational workflows.
The challenge is particularly acute for financial institutions that must balance the need for swift innovation with regulatory compliance. In Japan, where the banking sector is subject to periodic supervisory examinations, any lag in updating security standards could expose institutions to compliance risk. Tenon’s design therefore incorporates a governance layer that allows participating organisations to propose amendments to certification criteria, which are then reviewed by a consortium of stakeholders.
Nevertheless, the reliance on a consortium model raises questions about decision‑making speed and representation. Ensuring that Japanese interests are adequately reflected in the consortium’s deliberations will be essential if the network is to serve as a viable security backbone for Japan’s financial industry. Active participation by Japanese firms and possibly a government‑appointed observer could help align the network’s evolution with national policy objectives.
Future Outlook and Potential Expansion Beyond Finance
Although Tenon’s initial rollout targets the financial services sector, Liu indicated that the network is designed to be extensible to other industries that handle sensitive data, such as healthcare, logistics and public administration. The modular architecture permits the addition of sector‑specific verification criteria, meaning that Japanese firms in those domains could eventually adopt the same framework.
In the context of Japan’s aging population and the increasing reliance on AI‑driven health monitoring systems, a secure verification network could become a critical component of national health data governance. Similarly, the logistics sector, which is embracing AI for route optimisation and supply‑chain visibility, would benefit from assurance that AI agents are not compromised.
Looking ahead, the success of Tenon may influence how Japanese ministries, such as the Ministry of Health, Labour and Welfare (MHLW) and the Ministry of Land, Infrastructure, Transport and Tourism (MLIT), approach AI security in their regulatory frameworks. If the network demonstrates tangible risk reduction in its pilot phase, it could serve as a model for broader public‑sector adoption, reinforcing Japan’s position as a leader in secure AI deployment within the Asia‑Pacific region.
By Kenji Tanaka, Staff Writer
This article was produced with AI-assisted research and editorial support. Reporting is based on the source material cited below. Sources: CNA video report (19 November 2026); CNA; Global1.News
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Wow
0
Sad
0
Angry
0
Comments (0)