How to assess the ethics of information systems?

How should the ethics of information systems and algorithms be assessed? Jérôme Béranger addresses this fundamental question in DSIH (Décision Santé Informatique Hospitalière).

« The ethical audit of an algorithm cannot be improvised: it requires a rigorous method and objectifiable criteria. »

This article presents the ethical audit approach developed by GoodAlgo, built around several fundamental pillars:

• Transparency — are algorithmic decisions explainable and auditable?
• Non-discrimination — does the system avoid producing unjustified bias?
• Respect for privacy — is personal data processed in accordance with individuals' rights?
• Beneficence — is the system designed in the interest of end users?
• Accountability — who answers for decisions made or assisted by AI?

This evaluation grid, applicable to any information system incorporating algorithmic or AI components, forms the foundation of GoodAlgo's DEO (Digital Ethics Observatory) platform.