Your rights when an AI system decides
What the GDPR provides when a decision about you rests on automated processing: exceptions, information, human intervention, remedies.
On this page
In 30 seconds
- Where it happens
- The CNIL cites fields such as health, education, insurance, social protection and fraud detection[2], and also looks at tools that rank job applications automatically[3].
- What the law says
- As a rule, nobody has to accept a solely automated decision that produces legal or significant effects, except in three cases set out in Article 22[1].
- The reflex
- Ask for human intervention, in particular to obtain a review of your situation, express your point of view and contest the decision[2].
- If it has already happened
- The organisation informs the person of the action taken on their request within one month, extendable by two further months depending on complexity and the number of requests[1].
What it is
The GDPR gives every person the right not to be subject to a decision based solely on automated processing, including profiling, where that decision produces legal effects concerning them or similarly significantly affects them (Article 22)[1]. This right has exceptions, and safeguards come with them[2].
This page presents what the GDPR, the CNIL and Article 86 of the AI Act say, for an employee who is affected (by a CV screening, for example) or who receives a request from a customer or a colleague. It is not legal advice.
Straight answers
What is a fully automated decision?
For the CNIL, it is a decision taken about a person by algorithms applied to their personal data, with no human being involved in the process[2]. Article 22 targets those that produce legal effects or significantly affect the person[2].
A legal effect touches the person's rights and freedoms, for example the loss of a social benefit or of contractual rights[2]. A significant effect can exist without a legal effect: the CNIL cites a decision that has consequences for someone's financial situation or that leads to higher prices[2].
Profiling is a neighbouring but distinct notion: it is processing that uses an individual's data to analyse and predict their behaviour. An automated decision can do without profiling, such as automatic fines for traffic offences[2].
Is a fully automated decision always prohibited?
No. Paragraph 1 of Article 22 does not apply where the decision is necessary for entering into or performing a contract between the person and the controller, is authorised by Union or Member State law that lays down suitable safeguards, or is based on the person's explicit consent[1].
In the contract and consent cases, the controller puts suitable measures in place, at least the right to obtain human intervention, to express one's point of view and to contest the decision[1]. These decisions cannot be based on the special categories of data in Article 9, unless explicit consent or Article 9(2)(g) applies and safeguards are in place[1].
What must you be told about such a decision?
When data is collected from you, the controller must tell you that automated decision-making referred to in Article 22(1) and (4) exists and, at least in those cases, give meaningful information about the logic involved, as well as the significance and the envisaged consequences of the processing[1]. The same information is among what the right of access lets you obtain[1].
What is human intervention?
According to the CNIL, where no exception applies, human supervision that confirms or replaces the automated decision is necessary[4]. It raises the case of a CV-sorting algorithm that would put some CVs at the bottom of a queue too long for a human to check their relevance: consequences may exist in practice[4].
What does Article 86 of the AI Act add?
This article, titled "Right to explanation of individual decision-making" on the European Commission's website, covers any person subject to a decision taken by a deployer on the basis of the output of a high-risk AI system listed in Annex III (except point 2 of that annex). The decision must produce legal effects or similarly significantly affect the person, in a way they consider to have an adverse impact on their health, safety or fundamental rights[5]. The person then has the right to obtain from the deployer clear and meaningful explanations of the role of the AI system in the decision-making procedure and the main elements of the decision taken[5].
According to the CNIL, the rules on high-risk AI systems listed in Annex III apply from 2 December 2027[6].
The text sets two limits. It does not apply to uses for which Union or national law provides exceptions or restrictions, and it applies only to the extent that this right is not otherwise provided for under Union law[5].
The CNIL states that the AI regulation does not replace the requirements of the GDPR and is meant to complement them[6]. For an automatic CV-sorting system, it indicates that both texts apply[6].
Can you object to processing?
Article 21 gives a right to object, at any time, on grounds relating to the person's particular situation. It covers processing based on Article 6(1)(e) or (f), including profiling. The controller then stops the processing, unless it demonstrates compelling legitimate grounds that override the objection, or that the processing is needed to defend legal claims[1]. For direct marketing, including profiling related to it, objection is possible at any time[1].
How it works
The GDPR describes rights that work together: information about the existence of the automated decision, access to the data, then, in cases where an exception authorises the decision, human intervention, the chance to express one's point of view and the chance to contest[1]. According to the CNIL, for the exercise of certain rights, notably those of Article 22, precise explanations of the reasons for the decision are indispensable, and the complexity and opacity of some AI systems can make those explanations hard to provide[3].
In the credit field, the CNIL recalls that the judgments of the Court of Justice of the European Union in cases C-634/21 and C-203/22 confirm that scoring is an automated decision when it plays a determining role in granting credit, and establish a right to an explanation of how the mechanism that led to the decision works[7]. Its recommendation, published on 7 May 2026, sets out the conditions under which a credit decision may rest on fully automated processing, with three safeguards: transparency, human intervention and explainability[7]. It also anticipates that a legal authorisation for fully automated decisions on consumer credit will start to apply in November 2026[7].
This recommendation recalls the legal obligations and serves as a practical guide. The CNIL says it is possible to depart from it provided the choices are justified, in particular in the event of an inspection[7].
What you do
If you are affected, for example by a job application rejected without explanation:
- Check what the organisation told you. It had to inform you of the existence of an automated decision when your data was collected, with meaningful information about the logic involved[1].
- Use your right of access to obtain this information, as well as the purposes of the processing and the recipients of the data[1].
- Where the decision rests on a contract or on your explicit consent, ask for human intervention, for the chance to express your point of view and for the decision to be contested[1]. The CNIL describes this intervention as a way to obtain a review of one's situation and an explanation of the decision[2].
- If the organisation extends the one-month period, it must inform you of the extension and of the reasons within the first month[1].
- If the organisation does not act on your request, it must tell you why, and tell you that you can lodge a complaint with a supervisory authority and seek a judicial remedy[1].
- If the decision rests on a high-risk AI system listed in Annex III, Article 86 gives the right to ask the deployer for explanations, within the limits described above[5]. According to the CNIL, the rules on those systems apply from 2 December 2027[6].
If you receive a request from a customer or a colleague:
- The controller facilitates the exercise of the rights in Articles 15 to 22[1]. Information is provided free of charge, except for manifestly unfounded or excessive requests, which it is for the controller to demonstrate[1].
- Where there are reasonable doubts about who is making the request, the controller may ask for additional information to confirm their identity[1].
- The CNIL specifies that the organisation must explain how to exercise these rights[3].
Pitfalls
- A token human. For the CNIL, tools that automatically rank job applications can lead to a solely automated decision by design when applications are rejected, or pushed into a background tier that nobody checks for lack of time. It adds that, as a rule, the GDPR prohibits such practices, which are allowed only in exceptional conditions with safeguards[3].
- Decision support. The CNIL notes that, when the system only helps to decide, safeguards are also needed, notably regarding information[3].
- Correcting the data is not correcting the model. According to the CNIL, answering a request to rectify or erase training data does not necessarily mean rectifying or erasing the AI models built from that data[3].
Sources
- CNIL, “CHAPITRE III - Droits de la personne concernée”, 23 May 2016. https://www.cnil.fr/fr/reglement-europeen-protection-donnees/chapitre3 (accessed 14 November 2026)
- CNIL, “Profilage et décision entièrement automatisée”, 29 May 2018. https://www.cnil.fr/fr/profilage-et-decision-entierement-automatisee (accessed 14 November 2026)
- CNIL, “IA : comment être en conformité avec le RGPD ?”, 5 April 2022. https://www.cnil.fr/fr/intelligence-artificielle/ia-comment-etre-en-conformite-avec-le-rgpd (accessed 14 November 2026)
- CNIL, “Permettre le bon exercice de leurs droits par les personnes”, 28 March 2022. https://www.cnil.fr/fr/intelligence-artificielle/guide/permettre-le-bon-exercice-de-leurs-droits-par-les-personnes (accessed 14 November 2026)
- European Commission, “AI Act Service Desk - Article 86: Right to explanation of individual decision-making”. https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-86 (accessed 14 November 2026)
- CNIL, “Entrée en vigueur du règlement européen sur l’IA : les premières questions-réponses de la CNIL”, 12 July 2024. https://www.cnil.fr/fr/entree-en-vigueur-du-reglement-europeen-sur-lia-les-premieres-questions-reponses-de-la-cnil (accessed 14 November 2026)
- CNIL, “Octroi de crédit : la CNIL publie sa recommandation sur l’utilisation de données personnelles pour l’évaluation de la solvabilité”, 7 May 2026. https://www.cnil.fr/fr/recommandation-octroi-de-credit (accessed 14 November 2026)
Cite this article
Brain, “Your rights when an AI system decides”, Cyberpedia, 2026. https:/
BibTeX
@misc{brain_cp531_en_2026,
author = {Brain},
title = {Your rights when an AI system decides},
year = {2026},
howpublished = {Cyberpedia},
url = {https://startbrain.io/cyberpedia/en/articles/gdpr-automated-decisions-ai-rights/},
note = {CC BY 4.0}
}