Building the audit trail for the AI era
Human Oversight Labs was founded on a simple premise: as generative AI scales across workflows, probabilistic AI detectors will fail in a regulatory or legal setting.
Under emerging regulatory frameworks, most notably Article 50 of the EU AI Act, organisations, publishers and agencies need to adapt to new obligations regarding synthetic content transparency and human editorial control. As things stand, saying “a human reviewed this content” is no longer sufficient. Businesses now need verifiable, tamper-evident proof of human oversight.
Our mission is to build the definitive cryptographic record-keeping layer to bridge the gap between AI generation tools and editorial compliance, transforming AI from a liability to a genuine force multiplier.
Our approach: deterministic proof over AI guesswork
Unlike AI detectors, which can frequently produce false positives, our Art50 utility tool focuses on deterministic, verifiable attestation.
We don’t scan text to guess if an AI wrote it. Instead, we provide content teams with a structured verification workflow. When an editor fact-checks, edits, and approves an AI-assisted copy, our system logs verification parameters into a tamper evident hashed audit receipt.
This gives agencies, legal teams, and enterprise deployers a structured paper trail to demonstrate bona fide human editorial control.
Where we are today: the v1.0 manual utility
Version 1.0 of our platform is an early-stage Minimum Viable Product (MVP) released under a pre-incorporation beta license:
Manual Attestation Ledger: v1.0 operates as a structured record-keeping tool. It enables content teams to manually log review parameters and generate hashed digital audit receipts.
User-Declared Data Model: Receipts are compiled from self-reported reviewer inputs (such as completed compliance checklists and logged review durations).
Current Scope: At this early stage, the platform records what the user attests to, establishing a structured, tamper-evident paper trail for internal governance. It cannot yet guarantee an automatic exemption from disclosure obligations.
Where we are going: the v2.0 compliance engine
While v1.0 establishes the operational habit of logging human oversight, Version 2.0 will transition the platform form manual attestation to automated verification that can withstand regulatory scrutiny.
We are currently developing our v2.0 architechture to bridge the gap between internal workflows and compliance necessities by:
Automating Workflow Integration: Moving compliance into the background so teams can generate audit trails seamlessly within their existing work environments.
Providing Objective Verification Metrics: Replacing self-declared form fields with automated, undisputable, proofs of substantive human editorial effort.
Enhancing Cryptographic Rigor: Upgrading receipt infrastructure to ensure audit trails meet the highest standards of evidence required by EU regulatory bodies and legal counsel.
By evolving from a v1.0 manual attestation ledger into an automatic verification engine, we’re building the first long-term compliance asset uniquely designed to withstand the strictest regulatory scrutiny.