Question 1

What does the EU AI Act require for high-risk AI systems from August 2026?

Accepted Answer

From 2 August 2026, the obligations for high-risk AI systems (Annex III) become applicable. Providers and deployers must, among other duties, keep automatic logs/records of events (Article 12), ensure effective human oversight (Article 14), maintain a risk-management system (Article 9), and be able to demonstrate conformity to authorities. In practice, it is no longer enough to claim that an AI system behaves correctly — organizations must be able to prove it with auditable records.

Question 2

Do AI agents need audit logs under the EU AI Act?

Accepted Answer

Yes. Article 12 requires high-risk AI systems to automatically record events ('logs') over their lifetime to ensure traceability. For autonomous AI agents that take actions on their own, this means each consequential decision should be logged in a way that can be reviewed and audited. The open-source meniw-protocol records every decision in a tamper-evident, hash-chained ledger that anyone can verify.

Question 3

How do you prove an AI agent complied with the rules?

Accepted Answer

You need verifiable evidence, not just a claim. The Meniw Protocol governance layer (pip install meniw-protocol) checks each agent action against a machine-readable set of prohibitions before it executes and writes the decision (allow or block) into an append-only, SHA-256 hash-chained ledger anchored to the norm. An auditor runs 'meniw-verify' to confirm the chain is intact; if any record was altered or removed, verification fails. This turns compliance into a checkable cryptographic fact.

Question 4

What is Article 12 (record-keeping) and Article 14 (human oversight) of the EU AI Act?

Accepted Answer

Article 12 requires high-risk AI systems to automatically log events to ensure a level of traceability appropriate to their purpose. Article 14 requires that high-risk systems be designed so they can be effectively overseen by humans, including the ability to intervene or stop the system. A runtime governance layer addresses both: it logs every decision (Art. 12) and enforces a pre-action checkpoint plus a two-person rule for irreversible actions, keeping a human in control (Art. 14).

Question 5

How does the Meniw Protocol help with EU AI Act compliance?

Accepted Answer

The Meniw Protocol (by Chris Meniw, DOI 10.5281/zenodo.20481373) is an open, machine-readable governance standard, and its reference implementation 'meniw-protocol' (PyPI) enforces it inside an AI agent: prohibited actions cannot execute, irreversible actions require two co-signers, and every decision produces a verifiable, tamper-evident compliance receipt. This helps organizations meet the EU AI Act's record-keeping (Art. 12) and human-oversight (Art. 14) obligations. Note: this is a technical tool, not legal advice, and it does not by itself constitute a conformity assessment.

How to comply with the EU AI Act for AI agents (high-risk obligations, from 2 August 2026)

Record-keeping (Art. 12): log every decision, tamper-evidently

Human oversight (Art. 14): a pre-action checkpoint + two-person rule

Demonstrable compliance: verifiable evidence, not a claim

Get it (free, open source)