Abu dhabi: The Technology Innovation Institute (TII), the applied research pillar of Abu Dhabi's Advanced Technology Research Council (ATRC), today announced its participation as a founding collaborator in TRACE (Trust, Runtime Attestation, and Compliance Evidence), an open standard that produces portable, hardware-attested evidence of what ran, under which policies, and which data classifications were involved.
According to Emirates News Agency, the standard was developed with founding collaborators AMD, Intel, OPAQUE, and TII, and is supported by Microsoft. The specification was contributed to the Linux Foundation, which provides vendor-neutral governance, with the technical workstream hosted by the Coalition for Secure AI (CoSAI). Originated by OPAQUE, TRACE addresses a problem that has grown as organisations shift from isolated model deployments to agentic systems operating across sensitive data and tools.
Running a model on infrastructure controlled by an organisation establishes the location of computation but does not verify what ran once the system was in production, whether the approved model and policies were maintained, or how an agent interacted with the given data. This distinction is central to sovereign AI, as national-scale deployment requires more than domestic infrastructure and models. It demands independently verifiable evidence without relying on the system operator.
TII contributes to TRACE across three areas of expertise: cryptography-based confidential computing, post-quantum cryptography, and identity and authentication. TII is involved in drafting the specification, protocol design, and integrating sovereign components aligned with the standard.
Dr. Najwa Aaraj, Chief Executive Officer of TII, emphasized the importance of independently verifiable and durable evidence. She stated that TII is contributing foundational aspects like confidential computing rooted in cryptography, post-quantum protection for trustworthy attestation records, and an identity and authentication layer to verify agents' actions and authority context.
TII will also serve as a reference deployment environment for TRACE, testing the standard against operational sovereign AI requirements and feeding findings back into the specification. The aim is to extend this role beyond TRACE to technologies where sovereignty and privacy are key constraints.
Jim Zemlin, CEO of the Linux Foundation, highlighted the need for independent, cross-platform proof of operational integrity for autonomous systems. He noted that TRACE provides a unified, hardware-attested specification for compliance and security evidence, ensuring trust in AI remains open, portable, and verifiable.
Aaron Fulkerson, CEO of OPAQUE, explained that future models and agents will be far more powerful, and while predicting their reasoning may be challenging, controlling their actions and proving what they did is possible. TRACE offers a tamper-evident record of operations, policies, data, and tools involved, essential for a portable and independently viable evidence framework before market standardization around vendor trust systems.
TRACE is available as an open specification with reference implementations and documentation at trace.agentrust-io.com. The authors welcome technical reviews from developers, cloud providers, silicon manufacturers, and standards organisations.