Ethical Review Guide
A practical guide for businesses, government agencies, and organisations considering an AI Ethical Review with the Human Continuity Institute — what we assess against, what to submit, and what the review is (and is not).
What is an AI Ethical Review?
An AI Ethical Review is an independent expert assessment of an AI-related system, policy, or practice — conducted by our network of human specialists in ethics, law, technology, and human rights, with AI-assisted triage and a preliminary draft analysis to help our reviewers work efficiently. The final judgement is always made by a qualified human reviewer.
What guides our reviews
Our assessments are informed by leading international frameworks — not invented internally. The full, current list of sources we reference is maintained on our Knowledge Base page, and currently includes the EU AI Act and Ethics Guidelines for Trustworthy AI, the OECD AI Principles, and UNESCO’s Recommendation on the Ethics of Artificial Intelligence, alongside HCI’s own Constitution and Governance Charter.
Understanding AI risk categories under the EU AI Act
If your organisation operates in or serves the EU, the AI Act’s risk-based framework is a useful starting point for understanding what kind of review you may need:
- Prohibited practices — AI uses banned outright (e.g. social scoring, certain manipulative or biometric-categorisation systems). In force since February 2025.
- High-risk systems (Annex III — e.g. AI used in employment, credit scoring, law enforcement, critical infrastructure) — subject to conformity assessment, technical documentation, and risk-management obligations. Following the EU’s 2026 Digital Omnibus amendment, these obligations are now due 2 December 2027 for most standalone systems (previously 2 August 2026).
- Transparency obligations (Article 50 — e.g. disclosing AI-generated content, informing users they are interacting with an AI) — largely unaffected by the delay and still due 2 August 2026.
- Minimal-risk systems — most AI applications, with no mandatory obligations beyond voluntary codes of conduct.
Regulatory timelines change — this summary reflects our understanding as of publication. It is not legal advice; consult qualified counsel for your specific compliance obligations.
What you can submit
We can review a wide range of materials, submitted through our secure upload form. Examples include:
📄 System documentation
Technical documentation, model cards, architecture overviews, intended-use descriptions
⚖️ Governance & policy
AI use policies, risk-management plans, data governance policies, internal ethics guidelines
📊 Data & bias materials
Training data descriptions, bias/fairness testing reports, dataset documentation
👤 Human oversight
Human-in-the-loop procedures, escalation and appeal processes, incident response plans
📢 Transparency materials
User-facing disclosures, terms of use covering AI, consent language
📋 Compliance materials
Draft conformity assessment documentation, impact assessments (e.g. DPIA/FRIA), audit logs
Accepted formats: PDF, Word (DOC/DOCX), TXT, RTF, ODT. Max 25MB per file, 8 files / 60MB per submission.
What this review is not: HCI’s Ethical Review is an independent expert opinion, not a formal legal conformity assessment or certification under the EU AI Act or any other regulation. We are not a notified body. Organisations with mandatory conformity assessment obligations should also engage an accredited conformity assessment body and qualified legal counsel.
How the process works
- Submit — fill out the form on our AI Ethical Labs page, attaching relevant documents.
- AI-assisted triage — we confirm the submission is complete and reviewable.
- Preliminary analysis — AI drafts an initial assessment referencing the frameworks above, for our reviewers’ use.
- Human expert review — one of our specialists reviews the submission and the draft, and prepares your final report.
- Delivery — you receive a written assessment with findings and recommendations.