Abstract

Autonomous artificial intelligence agents interoperate in complex pipelines, delegating tasks and passing intermediate artifacts across boundaries. A hardware-attested inference identity framework provides verifiable cryptographic provenance for outputs produced by these artificial intelligence agent runtimes. An identity module anchored to a hardware root of trust computes a measurement set over the identity-defining configuration of the runtime. The configuration includes a model-weights digest, a system-prompt digest, and a guardrail-configuration digest. The module holds a binding key with a signing capability conditioned on current measurements matching the measurement set. For individual outputs, the module mints a signed inference certificate including the output digest, the measurement set, a monotonic counter value, and parent certificate digests of upstream artifacts incorporated into the output. A receiving runtime verifies the certificate chain to enforce admission policies.

Keywords: hardware attestation, cryptographic provenance, trusted execution environment, inference certificate, execution enclaves, multi-agent ecosystems.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS