EU Artificial Intelligence Act: human oversight, justice as high-risk, and the worker information right
Regulation (EU) 2024/1689 classifies AI systems intended to assist a judicial authority in researching and interpreting facts and law as high-risk, requires that such systems be designed so natural persons can effectively oversee them (Article 14), and gives workers' representatives a right to be informed before a high-risk system is used at the workplace (Article 26(7)). Its duties fall on providers and deployers; it confers no right on the person judged.
When this decides something for you
A reader in India or elsewhere is told that Europe has already regulated AI in courts, and asks what that regulation gives a litigant, a judge or an advocate. The answer is that it gives the litigant nothing directly, gives the judge a designed-in capacity to oversee and override, and gives workers a right to be told. Knowing the shape of the instrument prevents two errors: treating it as a model of what a litigant may demand, and treating it as irrelevant because it is not.
What the instrument is
Regulation (EU) 2024/1689, the Artificial Intelligence Act, published in the Official Journal on 12 July 2024. It is a product-safety style regulation: it classifies systems by risk, imposes design and documentation duties on providers, and imposes use duties on deployers. Its dates were altered by Regulation (EU) 2026/1744, the Digital Omnibus on AI, in force from 27 July 2026: high-risk obligations for standalone systems listed in Annex III now apply from 2 December 2027, and for AI embedded in Annex I products from 2 August 2028; the Article 5 prohibitions, in force since 2 February 2025, and the general application date of 2 August 2026 are unchanged.
What it provides
Justice as high-risk. Recital 61 states that "certain AI systems intended for the administration of justice and democratic processes should be classified as high-risk, considering their potentially significant impact on democracy, the rule of law, individual freedoms as well as the right to an effective remedy and to a fair trial", and that "it is appropriate to qualify as high-risk AI systems intended to be used by a judicial authority or on its behalf to assist judicial authorities in researching and interpreting facts and the law and in applying the law to a concrete set of facts", extending to alternative dispute resolution bodies where their outcomes produce legal effects. The corresponding entry is in Annex III.
Human oversight (Article 14). High-risk systems "shall be designed and developed in such a way, including with appropriate human-machine interface tools, that they can be effectively overseen by natural persons during the period in which they are in use" (14(1)). Oversight aims to prevent or minimise risks to health, safety or fundamental rights (14(2)). The persons to whom oversight is assigned must be enabled to understand the system's capacities and limitations, "to remain aware of the possible tendency of automatically relying or over-relying on the output produced by a high-risk AI system (automation bias)", to interpret its output correctly, "to decide, in any particular situation, not to use the high-risk AI system or to otherwise disregard, override or reverse the output", and to interrupt it (14(4)).
Worker information (Article 26(7)). Before putting a high-risk system into service at the workplace, deployers who are employers "shall inform workers' representatives and the affected workers that they will be subject to the use of the high-risk AI system", in accordance with Union and national law on worker information.
What it gives and what it does not
To the judge: a system built so that overriding it is possible, and an explicit warning against automation bias. To the worker: notice. To the provider and deployer: duties, enforced by market surveillance and penalties.
To the person judged: nothing in terms. The Act does not say that a litigant may demand a decision a human verified; it says the system must be overseeable. Oversight is a design property; verification is an act. A judge with a fully compliant Article 14 system who adopts its output unread has met the Regulation and failed the litigant. That distinction is the ground the canon's entitlement occupies, and it is why the entitlement cannot be read off the European instrument.
Article 26(7) is an information right, not a veto. It sits below co-determination and bargaining on the map of refusal capacity, and the Act itself provides (Article 2(11)) that it does not preclude national law or collective agreements more favourable to workers.
Where it stands in the record
The dates on this page follow the Omnibus as enacted; the recital and Article texts are quoted from the archived Official Journal copy. The Annex III entry itself was not read in that copy and is described, not quoted.
Provenance and method.
Where this was published
- Regulation (EU) 2024/1689 of 13 June 2024 (Artificial Intelligence Act)
- OJ L
- 12.7.2024
- recital 61
- Articles 14 and 26(7) (archived copy G1
- read 25 September 2026)
- Regulation (EU) 2026/1744 (Digital Omnibus on AI)
- OJ L
- 24.7.2026 (archived copy G2
- per the author's verified ledger of 28 July 2026)
Changelog
- 2026-09-25, version 1. First published.