§ AI Act BRIEFING

What the AI Act does not cover — and the four edges where the exemptions are narrower than they look.

Article 2 of Regulation (EU) 2024/1689 carves out four scope exemptions and one fine-relief mechanism. Each has a sharp edge that catches teams who read the headline and stopped there.

Summary

The AI Act's Article 2 looks generous on first reading. It excludes military AI, scientific R&D, personal use and pre-market testing. Open-source GPAI models get partial relief under Article 53(2). SMEs are not exempt but get fine proportionality under Article 99(6).

Each of these has a narrow edge. A defence contractor using an AI for HR alongside the military system is in scope. An open-source model crossing the systemic-risk threshold loses the carve-out. A research lab putting a model on Hugging Face has placed it on the market.

The exemption text is short. The case law that interprets it does not exist yet. Below is what the regulation actually says, what it does not say, and what Fontvera flags as the four exposure edges.

Who this applies to
Any team building, distributing or deploying AI in the EU that is currently relying on a scope exemption under Article 2 — defence contractors, research institutes, open-source GPAI publishers, SMEs and individual developers.
Compliance deadline
High-risk obligations: 2 August 2026 as written; provisionally 2 December 2027 for Annex III under the Digital Omnibus agreement of 7 May 2026, pending formal adoption.
§ Key articles

What the law says

AI Act Article 2(3)
Exclusion for AI exclusively for military, defence or national security purposes.
AI Act Article 2(6)
Exclusion for AI specifically developed and put into service for the sole purpose of scientific research and development.
AI Act Article 2(8)
Exclusion for any research, testing or development activity prior to placing on the market.
AI Act Article 2(10)
Exclusion for natural persons using AI in the course of purely personal non-professional activity.
AI Act Article 53(2)
Open-source GPAI carve-out from some transparency obligations — does not extend to systemic-risk models.
AI Act Article 99(6)
Authorities must give due regard to undertaking size and turnover when setting fines.
§ Detail

In depth

1. Military, defence and national security — Article 2(3)

The AI Act does not apply to AI systems placed on the market, put into service, or used "exclusively for military, defence or national security purposes" — irrespective of the entity carrying out those activities. Source: Article 2(3), Regulation (EU) 2024/1689.

The edge: the word is exclusively. Dual-use AI is not exempt. A computer-vision model trained for ISR (intelligence, surveillance, reconnaissance) and also licensed to a port authority for cargo screening loses the exemption for the civilian deployment. The defence contractor remains an AI Act provider for the civilian variant. The same model, two regimes.

Knock-on: AI used by national security agencies in border control, law enforcement intelligence and counter-terrorism may sit in a grey zone. Recital 24 stresses that the exemption does not extend to AI used for general-purpose policing, immigration controls or border management — those fall under Annex III high-risk categories. The AI Office is expected to publish further guidance, but pending that, the operator carries the classification risk.

2. Scientific research and development — Article 2(6) and 2(8)

Two distinct exemptions sit here. Article 2(6) excludes AI systems "specifically developed and put into service for the sole purpose of scientific research and development." Article 2(8) excludes any research, testing or development activity carried out before the AI system is placed on the market or put into service, except real-world testing.

The edge: the moment the model is placed on the market — including being made available "for distribution or use on the Union market in the course of a commercial activity, whether in return for payment or free of charge" (Article 3(9)) — both carve-outs end. Publishing model weights on Hugging Face counts as placing on the market. So does a public API endpoint, even an unmonetised one.

Real-world testing outside the regulatory sandbox regime (Articles 57–63) is explicitly not covered by Article 2(8). A research team running a clinical pilot of a high-risk medical AI in a hospital is in scope from day one of testing — not from the day a paper is published.

3. Personal, non-professional use — Article 2(10)

"This Regulation does not apply to obligations of deployers who are natural persons using AI systems in the course of a purely personal non-professional activity." Source: Article 2(10), Regulation (EU) 2024/1689.

The edge: the exemption only covers the deployer. The provider obligations on the AI vendor (transparency under Article 50, conformity if high-risk, GPAI obligations under Article 53) remain. A consumer using a deepfake app for a private prank is not the regulated party — the app vendor still is.

Side-channel: "purely personal" is narrower than "non-commercial." A freelance journalist using an AI tool for paid client work is not personal use. A teacher using a marking AI on student work is professional. A lawyer redlining a contract with an AI is professional. The carve-out does not protect side projects that touch professional obligations.

4. Open-source GPAI — Article 53(2), with a hard ceiling

Open-source general-purpose AI providers are exempt from the technical documentation and downstream-information obligations of Article 53(1)(a) and (b), provided the model is released under a free and open-source licence with weights and architecture publicly available. The copyright policy under Article 53(1)(c) and the training data summary under Article 53(1)(d) remain mandatory. Source: Article 53(2), Regulation (EU) 2024/1689.

The edge — the systemic-risk ceiling: the open-source carve-out evaporates the moment the model is classified as "GPAI with systemic risk" under Article 51. The threshold is cumulative training compute exceeding 10²⁵ FLOPS, or a Commission decision based on capability and reach. A 50B-parameter open-weights model trained at frontier scale is not exempt — every Article 55 obligation (model evaluation, adversarial testing, systemic risk mitigation, cybersecurity, incident reporting) applies.

And a downstream catch: open-source carve-outs at provider level do not modify the obligations of deployers. A bank using an open-source GPAI for credit scoring is in Annex III 5(b) territory regardless of what the model's licence says.

5. SMEs — Article 99(6) is fine relief, not exemption

There is no SME exemption from substantive AI Act obligations. Article 99(6) requires market surveillance authorities to give due regard to the size of the undertaking and its annual turnover when imposing administrative fines. SMEs and start-ups also benefit from priority access to regulatory sandboxes (Article 57) and simplified documentation pathways "where possible." The conformity assessment, risk management, post-market monitoring and incident reporting obligations are unchanged.

What the AI Act explicitly does not exempt

Real numbers Fontvera tracks

Penalty exposure if you misclassify

Wrongly relying on an exemption is treated as the underlying breach, not as a separate offence. If a system you classified as exempt turns out to fall under Article 5 prohibition, the ceiling is €35,000,000 or 7% of worldwide turnover (Article 99(1)). For misclassified high-risk systems the ceiling is €15,000,000 or 3% under Article 99(2). For supplying inaccurate information to authorities while defending an exemption claim, Article 99(3) adds up to €7,500,000 or 1%.

Run your free AI Act compliance diagnostic

If you are inside an exemption edge — open-source GPAI, dual-use defence, hospital research pilot — Fontvera's diagnostic returns the classification with the specific articles that apply.

→ Run the AI Act diagnostic

§ Action items

Practical steps

01
Audit any system tagged 'exempt' against the four-edge test: dual-use, market placement, professional context, systemic-risk threshold.
02
Document the exemption claim in writing with the specific Article 2 paragraph and the reason. The burden of proving an exemption sits on the operator.
03
For open-source GPAI: track training compute. Crossing 10²⁵ FLOPS dissolves Article 53(2) and triggers Article 55 systemic-risk obligations.
04
If the system is in research pilot today and a public release is planned, run the conformity assessment in parallel — placing on the market is the trigger, not commercial launch.
§ What Fontvera found

Documents in our corpus

eiopa EU Fetched 2026-04
Opinion on Artificial Intelligence governance and risk management
eurlex EU Fetched 2026-04
EUR-Lex: 32025R0454 (2025-03-07)
ai_office EU Fetched 2026-06
§ Cross-references

Related Fontvera intelligence

Need a cross-border briefing on this?
Search Fontvera ↵ Run the AI Act diagnostic
AI Act Article 50 transparency
50 days
until 2026-08-02, when Article 50 transparency obligations apply (unchanged). Annex III high-risk obligations move provisionally to 2 December 2027 under the Digital Omnibus agreement of 7 May 2026, pending formal adoption.
Preparing for 2 August 2026? Read the EU AI Act August 2026 deadline requirements checklist.