ALTAI

A self-assessment checklist from the EU High-Level Expert Group on AI that helps you test an AI system against the seven requirements for trustworthy AI and spot where you need extra measures.

What it is¶

ALTAI turns the 2019 Ethics Guidelines for Trustworthy AI into questions, grouped under seven requirements: human agency and oversight, technical robustness and safety, privacy and data governance, transparency, diversity, non-discrimination and fairness, societal and environmental well-being, and accountability. You answer them with a multidisciplinary team, ideally after a fundamental rights impact assessment, and pick the questions that fit your system and sector. It is a reflection tool, not a certification: completing it does not prove compliance. It comes as a PDF and as an interactive web version with extra explanatory notes.

In healthcare¶

ALTAI is sector-neutral. In clinical AI it works well as a broad first pass over ethical and technical risks before you move to medical-device and AI Act specific checks.

Where to use it in The ELSA Way¶

• Risk management planning › 1. Risk identification: Its seven requirement areas give your interdisciplinary team a broad prompt list, so you catch ethical, societal and technical risks that a purely technical review would miss.

• Ethical review and approval › Application-specific ethical issues: Many of its questions, on human oversight, non-discrimination, accountability and environmental well-being, line up with the ethical concerns you need to document during design.

References¶

• High-Level Expert Group on Artificial Intelligence. Assessment List for Trustworthy Artificial Intelligence (ALTAI) for self-assessment. Brussels: European Commission; 2020. undefined (2020)

Related tools¶

• IAMA: A structured discussion guide that helps a public organisation decide whether and how to use an algorithm, with a step-by-step check of its effects on fundamental rights.

• AI Impact Assessment: A question set that walks a multidisciplinary team through the purpose, impact, technical design, data and risks of an AI system, and records the decision whether to use it.

• Algoritmekader: A Dutch government website that brings together the legal requirements, recommended measures and supporting tools for responsible use of algorithms and AI, organised by lifecycle phase, role and topic.