The EU withdrew its AI Liability Directive, so there is no AI specific fault regime for a robot that harms someone
The proposal to adapt non contractual civil liability rules to artificial intelligence, COM(2022)496, appears in Annex IV of the Commission work programme for 2025 as a withdrawal, and the withdrawal was published in the Official Journal on 6 October 2025. Nothing replaced it. A victim of a robot falls back on the product liability regime and on national fault law.
The verdict
Verified
The document exists. The ledger fetched it at its publisher and quotes it.
Key facts
What the sources say
- Record ID
- PAG-2026-0005
- Kind
- Binding law
- Jurisdiction
- European Union
- Last verified
- Added
- Official Journal C/2025/5423 of 6 October 2025 lists COM(2022)496 final, procedure 2022/0303 (COD), among the withdrawn Commission proposals.
- Commission work programme 2025, Annex IV item 32, gives the reason: no foreseeable agreement, and the Commission will assess whether another proposal should be tabled or another approach chosen.
- The proposal would have eased the burden of proof and given disclosure of evidence about high risk AI systems; neither survives in binding EU law.
Dimension by dimension
2 dimensions, each one stated, silent or open
Liability, The AI inside it. Stated means the document you can open below says it; silent means the ledger read the document and it does not.
- LiabilityStated
- No AI specific presumption of causation exists in EU law; the claimant proves the case under the product liability regime or national fault rules.EUR-Lex, Official Journal C/2025/5423, withdrawal of Commission proposals, primary source, 6 October 2025.
- The AI inside itStated
- The withdrawal leaves the opacity problem to the courts, since the disclosure duty that was meant to answer it was in the withdrawn text.EUR-Lex, Official Journal C/2025/5423, withdrawal of Commission proposals, primary source, 6 October 2025.
What it changes
For a robot maker or a deployer
Do not build a compliance plan around an AI liability regime that does not exist. Two instruments carry the risk instead: the revised Product Liability Directive for defective products, and the national fault law of every member state you sell into, which differs country by country. The practical consequence is that your evidence trail, not a harmonised presumption, decides how a claim goes.
Sources
What this record was verified against
- EUR-Lex, Official Journal C/2025/5423, withdrawal of Commission proposalsPrimary · 6 October 2025
- EUR-Lex, Commission work programme 2025, COM(2025) 45 final, Annex IVPrimary · 11 February 2025
Related
Records that sit beside this one
The revised EU product liability rules make software a product and reach a robot that changes after it is sold
European Union · verified 15 September 2026
Article 4(1): product means all movables, even if integrated into or inter connected with another movable or an immovable, and it includes electricity, digital manufacturing files, raw materials and software.
Open question: who is liable when a behaviour the robot learned causes the harm?
Global · verified 15 September 2026
The AI Liability Directive proposal, which would have eased proof for exactly this case, was withdrawn and published as withdrawn in the Official Journal on 6 October 2025.
Open question: is a fleet wide behaviour update a new placing on the market?
European Union · verified 15 September 2026
The Machinery Regulation applies to machinery placed on the market from 14 January 2027 and treats a substantially modified machine as a new product.
Open question: is a humanoid in a home a machine, a consumer product or a care robot?
European Union · verified 15 September 2026
The General Product Safety Regulation applies only where no specific Union law with the same objective regulates the safety of the product, and only to risks those rules do not cover.
The EU has no dedicated robotics strategy document, only a research policy page and a section of the AI strategy
European Union · verified 15 September 2026
The Commission robotics policy page describes a portfolio of over 120 research projects and coordination actions and the SPARC public private partnership.
The EU's robotics plan is one section of the Apply AI Strategy, and it funds uptake rather than rules
European Union · verified 15 September 2026
Section 2.2 opens on the installed base: more than 90.000 industrial robots installed in 2023 and more than 400 service robotics producers in Europe.
Cite this record
Free to reuse under CC BY 4.0, with attribution. The record ID PAG-2026-0005 is permanent and is never reused.
- In a sentence
- According to the GAGE Physical AI Governance Atlas (as of 15 September 2026), the eu withdrew its ai liability directive, so there is no ai specific fault regime for a robot that harms someone.
- APA
- GAGE (Global Academy of Generative-AI Education). (2026). The EU withdrew its AI Liability Directive, so there is no AI specific fault regime for a robot that harms someone. Physical AI Governance Atlas. Retrieved 15 September 2026, from https://www.gage.academy/tools/physical-ai-governance-atlas/records/PAG-2026-0005-eu-ai-liability-directive-withdrawn
- MLA
- "The EU withdrew its AI Liability Directive, so there is no AI specific fault regime for a robot that harms someone." Physical AI Governance Atlas, GAGE (Global Academy of Generative-AI Education), 15 September 2026, https://www.gage.academy/tools/physical-ai-governance-atlas/records/PAG-2026-0005-eu-ai-liability-directive-withdrawn.
- Chicago
- GAGE (Global Academy of Generative-AI Education). "The EU withdrew its AI Liability Directive, so there is no AI specific fault regime for a robot that harms someone." Physical AI Governance Atlas. Last modified 15 September 2026. https://www.gage.academy/tools/physical-ai-governance-atlas/records/PAG-2026-0005-eu-ai-liability-directive-withdrawn.
- Permalink
- https://www.gage.academy/tools/physical-ai-governance-atlas/records/PAG-2026-0005-eu-ai-liability-directive-withdrawn
Last updated . Every record re verified . The ledger is checked monthly, first Monday, and the same day for any new law or standard publication.
Back to the full ledger, or every record for European Union and every binding law record.
GAGE briefings tell you which AI regulation deadlines are coming, what they actually require of you, and when a program opens.