Bern
Lisbon
New York
Key articles on responsible AI for teams that build, integrate or use it. Written by AIEI, open to everyone.
Press / to search from anywhere on the page
What can AI do, and where does it fall short? Why AI output can be wrong or inconsistent, and a simple rule for when a person must check it.
What can AI do, and where does it fall short? Why AI output can be wrong or inconsistent, and a simple rule for when a person must check it.
AI governance glossary: plain-English definitions of key AI Act, GDPR and AI risk terms, why each matters for a company, and where to read more.
AI ethics sets the principles, governance sets roles and decisions, compliance meets specific rules. How they differ and why a company needs all three.
An AI inventory lists every AI tool your company uses, including shadow AI. Where to look, a common mistake to avoid, and a ready-to-use template.
Who should be responsible for AI in a small company? Often existing roles: one programme owner, an owner for each use case and experts when needed.
How to approve a new AI use case: request, risk triage, testing, pilot and a decision by a named owner, with a fast-track rule and a request form.
What an AI use policy should include: permitted uses, data rules, tool approval, human review, incidents, training and ownership, and the mistake to avoid.
Which AI documentation to keep: inventory, assessments, approvals, tests, training, vendor files, incidents and changes, plus AI Act and GDPR record rules.
How to monitor AI after launch: what to track, when to reassess a use case, and when monitoring becomes a legal duty under the EU AI Act.
Human oversight of AI works only if reviewers have time, information and authority. What GDPR and the EU AI Act require, and how to set it up.
The questions to ask an AI vendor about your data, model changes, incidents, output rights and its EU AI Act role, with a ready-to-send questionnaire.
What AI transparency requires: EU AI Act Article 50 disclosures, duties to explain decisions under the GDPR and Article 86, and sample notices.
Which AI laws apply depends on where your customers are and where AI output is used, not only where you are registered. Three questions to check.
Do you need ISO 42001 certification? What the standard is, how it differs from the law and other AI frameworks, and when certification makes sense.
What the EU AI Act requires: four risk levels, rules for general-purpose AI, which dates already apply and what the 2026 Digital Omnibus changed.
Provider or deployer? How the EU AI Act assigns roles, when a deployer becomes a provider under Art. 25, and a short decision tree to find your role.
Where AI bias comes from, where it matters most in hiring, lending, pricing and customer service, and how to screen for it without data scientists.
What to do in an AI incident: six response steps, and the GDPR and EU AI Act reporting duties that may apply, including who must meet them.
AI and copyright raise questions about what goes into a tool and what comes out. What to check in vendor terms and before you publish AI content.
When not to use AI: the EU AI Act's prohibited practices, three signals that a use case should be paused, and why saying no is a valid outcome.
Whether you can share data with AI tools depends on the data and the tool's terms. A traffic-light rule, plan checks and the GDPR questions that remain.
What AI literacy training staff need under Art. 4 of the EU AI Act: proportionate measures by role, from all staff to leadership, and how to record them.
No articles match “”
Try a shorter phrase, or check the glossary for the term you need.
Show all articles
Join us as a member and play a vital role in shaping a future where AI is created responsibly, with integrity, transparency, and fairness at its core.