EU AI Act Article 5 — Prohibited practices

Note: this explainer is informational and framework references are never a compliance guarantee — align your program with your legal counsel.

What Article 5 requires

Article 5 prohibits AI practices the EU deems unacceptable — among them manipulative techniques that materially distort behavior and cause harm, exploitation of vulnerabilities, social scoring, and untargeted scraping for facial-recognition databases. These prohibitions are already in force, with the AI Act's highest penalty band attached.

What it means for your company

No sane company sets out to deploy a prohibited system — the real risk is unknowingly using one: an employee adopts an unvetted tool whose functionality crosses an Article 5 line, and your organization is the deployer. The practical duty is therefore control over which AI tools enter your environment at all.

How Shield operationalizes it

The Block unsanctioned / high-risk AI tools preset (referenced to Art. 5) fully blocks known high-risk tools, org-wide, for whichever of them appear in your environment. Upstream of that, AI Discovery makes every new tool visible for triage — you decide what is sanctioned before it becomes habit, and the "Ungoverned" metric shows your open exposure at a glance.

Apply it via the preset library; it is part of the EU AI Act & GDPR Starter bundle.

The outcome

Prohibited-class and high-risk tools cannot quietly take root: discovery surfaces them, the block policy stops them, and the audit trail evidences that your organization exercised the control Article 5 assumes.