Blog 1 August 2026 · 4 min
AI regulation 2026: the right to an explanation of why your CV was rejected
Thousands of applications vanish into sorting algorithms every day without any explanation ever being given. With the European AI Act coming into force in 2026, candidates now have an unprecedented lever: the right to an explanation of automated decisions. According to stubby-ai.com, knowing this right is essential, but relying on it alone is not enough, because it always applies too late, after the rejection. In this article, you will learn what the regulation actually says, how to word a request for an explanation to a recruiter, and above all how to audit your application beforehand so you never have to demand answers.
Your CV used to vanish into thin air. In 2026, that changes.
You applied. You never heard back. No explanation, no acknowledgement from an actual person.
For years, the ATS (automated applicant tracking software) worked like a black box. You had to accept its verdict without ever being able to challenge it.
The good news: with AI regulation in 2026, the balance of power is shifting. Candidates gain a little-known lever: the right to an explanation of automated screening decisions.
I'm Albane, and I'm going to tell you frankly what this right really changes. And above all, what it doesn't.
What the regulation actually says
The European AI Act classifies recruitment as a "high-risk" use. That makes sense: screening candidates has a direct impact on their working lives.
As a result, three key principles apply to companies that use AI for screening.
1. Transparency. From 2 August 2026, the transparency obligations in Article 50 come into effect. The European Commission explains this in its official guidelines. In practice: you must be able to know that an AI has analysed your application.
2. Non-discrimination. A scoring system must not reproduce biases linked to gender, age or origin. This is a central requirement of the text.
3. Explainability. This is the heart of the matter. Article 86 of the AI Act provides a right to a clear and meaningful explanation when a high-risk automated decision has a significant effect on you.
You can ask to understand the role the AI played and the main factors that led to the decision.
Now, let's be honest. This right does not give you access to the source code or the full details of the algorithm.
There are also exceptions provided for by EU or national law. What is an enforceable right on paper can sometimes be vague in practice.
The fairest view: see this right as a lever for transparency and challenge, not as a guarantee of overturning a rejection.
The paradox you absolutely need to understand
Here is the trap I don't want you to fall into.
Having a "right to an explanation" is useless if your CV is eliminated before it even reaches a human. One blunt semantic filter, and it's over.
As Reuters points out, scoring and pre-screening tools can filter out an application even when a human still makes the final decision. Legal compliance does not prevent this initial filtering.
In other words: legal transparency never replaces optimising beforehand. Asking for an explanation after a rejection means arriving too late.
The well-informed candidate does both: they know their rights and they bulletproof their application so they never need to invoke them.
The practical method
Let's get concrete. Here is how to take back control today.
Spotting whether a company uses AI screening
A few reliable signs:
- Mentions of GDPR or AI in the application process or privacy policy.
- Well-known recruitment platforms with built-in automated scoring.
- Highly structured forms asking for specific keywords rather than a free-form CV.
- A screening questionnaire before any human contact.
Writing a request for an explanation
If you receive a rejection and suspect an automated decision, you can write to the company. Contact its HR department or its data protection officer (DPO).
Here is a simple, polite template.
Subject: Request for an explanation regarding the processing of my application
Dear Sir or Madam,
I applied for the position of [job title] on [date], reference [no.].
Under my rights relating to automated decisions (AI Act,
Article 86, and GDPR), I would like to know whether an AI system
was used to analyse my application.
If so, I would be grateful if you could tell me the main criteria
that led to the decision, as well as the role played by this system.
I remain at your disposal and thank you in advance for your reply.
Kind regards,
[Full name, contact details]
Expect a response time of around one month, as with a standard access request. Stay factual and polite: the aim is to get information, not to start a dispute.
Anticipate rather than endure: generate your own explainability report
This is where it all happens. Rather than waiting on a recruiter's goodwill, audit your own application before you apply.
The idea is simple: produce your own explainability report. A compatibility score between your CV and the job ad, missing keywords, concrete areas for improvement.
A general-purpose tool such as ChatGPT quickly shows its limits for this specific task, as I explained in this article on the AI that admits its own limits. A specialised AI, trained for CV-to-job matching, is far sharper.
That is exactly the thinking behind Stubby: cross-checking your CV against the job ad to tell you, before you send it, what's not working. You become your own auditor.
And to go further on building your CV, have a look at my method for beating the productivity filter that eliminates 90% of CVs. It's the perfect complement to this anticipation step.
In 2026, you don't wait, you audit
Let's sum up plainly.
- The right to an explanation exists, but it is not a magic wand.
- It still depends on exceptions and how it is applied.
- It comes after the rejection, so too late to change the outcome.
- Anticipation, on the other hand, works before you send.
Transparency is a right, and that's good news. But anticipation remains your best weapon.
Thanks to AI regulation in 2026, you can now demand answers. Thanks to a serious self-assessment, you will almost never need to. The winning candidate no longer endures the algorithm: they audit it themselves.
« "Transparency is a hard-won right; anticipation is a weapon no one can take away from you." »
| Section | Key messages |
|---|---|
| The new right | The 2026 AI Act classifies recruitment as a high-risk use and establishes a right to an explanation of automated decisions (Article 86). |
| The 3 pillars | Transparency about the use of AI, non-discrimination, explainability of criteria. Applicable from 2 August 2026. |
| The limits | No access to source code, exceptions possible. A lever for challenge, not a guarantee of success. |
| The paradox | The right is exercised after the rejection. If the CV is filtered out before any human sees it, the explanation comes too late. |
| The method | Spot AI screening (GDPR/AI mentions), send a template email to the DPO/HR, and above all self-assess your application beforehand. |
| Stubby's approach | Generate your own explainability report (matching score, missing keywords) rather than depending on the recruiter. |
