Login Get a Demo
01 Glossary · Compliance

What Are Prohibited AI Practices?

The Article 5 list in full, including the two prohibitions added in 2026, with the dates each applies and the fine tier that backs them.

The short answer

What are prohibited AI practices?

Last updated
Read time
4 min read
From
Cranium AI
Also called
Unacceptable-risk AI, Article 5 prohibitions
Share on LinkedIn

Prohibited AI practices are the uses of AI that Article 5 of the EU AI Act bans outright.

The list has ten lettered points: eight original prohibitions that have applied since February 2, 2025, and two added in 2026 that apply from December 2, 2026. Breaches carry the Act’s highest fines.

Key takeaways
  • Most points prohibit placing the system on the market, putting it into service and using it.
  • The eight original prohibitions have applied since February 2, 2025.
  • Two points added in 2026, on non-consensual intimate imagery and child sexual abuse material, apply from December 2, 2026.
  • Fines reach EUR 35 million or 7% of worldwide annual turnover, whichever is higher (Article 99(3)).
02 The list

What does Article 5 prohibit?

PointProhibited practice
(a)AI that deploys subliminal, purposefully manipulative or deceptive techniques that materially distort behavior and cause, or are reasonably likely to cause, significant harm
(b)AI that exploits vulnerabilities due to age, disability or a specific social or economic situation to materially distort behavior in a way that causes, or is reasonably likely to cause, significant harm
(ba)AI that generates or manipulates realistic images, video, audio or similar material of an identifiable person’s intimate parts or sexually explicit activity without that person’s explicit consent (added 2026)
(bb)AI that generates or manipulates child sexual abuse material within the meaning of Directive 2011/93/EU (added 2026)
(c)Social scoring that leads to detrimental treatment in unrelated contexts, or treatment that is unjustified or disproportionate
(d)Predicting the risk of a person committing a criminal offense based solely on profiling or personality traits
(e)Creating or expanding facial recognition databases through untargeted scraping of facial images from the internet or CCTV footage
(f)Inferring emotions in the workplace and in education institutions, except for medical or safety reasons
(g)Biometric categorization that deduces race, political opinions, trade union membership, religious or philosophical beliefs, sex life or sexual orientation
(h)Real-time remote biometric identification in publicly accessible spaces for law enforcement, unless strictly necessary for one of three listed objectives
03 The limits

Where do the prohibitions stop?

01

Exceptions sit inside the points

Emotion inference is allowed for medical or safety reasons, and criminal risk assessment may support a human assessment already based on objective, verifiable facts linked to a criminal activity.

02

The 2026 additions are scoped

Under Article 5(1a), placing such a system on the market is prohibited only where generating the material is its intended purpose, or a reasonably foreseeable and reproducible outcome that it lacks adequate safeguards to prevent.

03

Guidelines explain, courts decide

The Commission published guidelines on prohibited practices on February 4, 2025. They are non-binding, and authoritative interpretation is reserved for the Court of Justice of the European Union.

Commission guidelines
04 In practice

How to screen your own AI use

  1. Inventory every AI use

    List the AI systems you build, buy and use, including AI features inside vendor software.

  2. Compare with the ten points

    Look hardest at emotion inference at work, biometric categorization, scoring of individuals and generative tools that can produce intimate imagery.

  3. Record the outcome

    Keep the screening result with the inventory entry, and repeat it when a system’s purpose or features change.

05 Keep going

Related reading.

06 Questions

Frequently asked questions.

Straight answers, kept current. Last updated .

Ask us anything →
When did the prohibitions start to apply?

The eight original prohibitions have applied since February 2, 2025. Regulation (EU) 2026/1744 added points (ba) and (bb) and paragraphs 1a and 1b, which apply from December 2, 2026 under Article 113 as amended. The Commission’s implementation timeline dates national and EU-level enforcement from August 2, 2026.

What is the fine for a prohibited practice?

Article 99(3) sets administrative fines of up to EUR 35 million or, if the offender is an undertaking, up to 7% of its total worldwide annual turnover, whichever is higher. For SMEs, including start-ups, the lower of the two applies under Article 99(6).

Do the prohibitions bind deployers or only providers?

Both. Most points prohibit placing on the market, putting into service and use, so a company using a banned system is exposed alongside its supplier. For the two points added in 2026, use is prohibited where the deployer uses the system to generate or manipulate the material.

Is emotion recognition banned everywhere?

No. Article 5(1)(f) prohibits AI that infers emotions in the workplace and in education institutions, except for medical or safety reasons. Elsewhere, emotion recognition is listed in Annex III as a high-risk use, and Article 50 requires deployers to inform the people exposed to it. This page is general information, not legal advice.

07 See it for yourself

From the definition
to the evidence.

See how Cranium AI discovers, observes, governs, secures and proves every AI model, agent and vendor in your enterprise.