Alethexis
  • Pricing
  • Trust

Stay ahead of the AI Act

Regulatory analysis, AI news and product updates. No noise.

I agree to receive communications from Alethexis: content about AI and regulation, and product news. I can unsubscribe at any time.

Data protection — newsletter

Controller: ALETHEXIS, S.L. (NIF B88758057). Purpose: to send you the Alethexis newsletter (content about AI and regulation, and product news). Legal basis: your consent (Art. 6(1)(a) GDPR), which you can withdraw at any time. Recipients and transfers: we use providers acting on behalf of Alethexis for hosting, security and email delivery. No disclosure of your data to third parties for their own purposes is planned, except where required by law. Some providers may process data outside the EEA with the safeguards set out in the additional information. Rights of access, rectification, erasure, objection, restriction and portability: privacidad@alethexis.com. You may lodge a complaint with the Spanish Data Protection Authority (AEPD, www.aepd.es) or with the supervisory authority of your Member State (Article 77 GDPR). More information in the privacy policy.

Alethexis
Take the AI Act assessment (3 minutes)14-day free trial

Product

  • M1 Visibility
  • M2 Governance
  • M3 High-Risk Ready
  • Agentic Add-on
  • How it works
  • All features
  • Pricing
  • AI governance software

Solutions

  • For DPOs and consultants
  • Shadow AI
  • If you develop AI

Sectors

  • Dental Clinics
  • B2B Distribution
  • Food Distribution
  • Recruitment and HR
  • Private Education
  • Private Healthcare
  • Financial Services
  • Logistics and Transport
  • Professional Services and Consulting
  • Other sectors
  • All sectors

Resources

  • The Route: AI governance in 30 days
  • AI Act assessment
  • Templates
  • EU AI Act guide
  • Glossary
  • AI Act timeline
  • Blog
  • Changelog

Trust

  • Overview
  • Trust Center
  • Security
  • Accessibility

Company

  • About us
  • Contact
  • hello@alethexis.com
  • Partners

Legal

  • Legal Notice
  • Privacy Policy
  • Terms of Service
  • DPA (Data Processing Agreement)
  • Cookie Policy
  • Subprocessors & Service Providers Register
© 2026 Alethexis, S.L. · NIF B88758057 · C/ Tirso de Molina 36, 08940 Cornellà de Llobregat (Barcelona) · The enterprise standard. Built for SMEs.

Alethexis structures the governance, risk, operation and evidence of AI use. Using the platform does not constitute legal advice, nor does it by itself establish regulatory compliance.

LinkedIn · Contact · Product data in the EU 🇪🇺 · See the subprocessors & service providers register
By Rafael Luque Ocaña·23 July 2026

Biometrics at work: verification isn't identification, and the difference changes everything

Clocking in with a fingerprint, opening doors with your face: workplace biometrics is best analysed through one distinction that orders everything else — one-to-one verification versus one-to-many identification — and the necessity and proportionality test that the impact assessment requires.

Few everyday technologies pack in as much regulatory density as a fingerprint reader on a company door. It looks like a clocking-in formality; it's processing of biometric data — a special category under Article 9 GDPR when used to uniquely identify a person — over the workforce, and under an AI Regulation that reserves some of its harshest boxes for biometrics. And yet the whole analysis falls into place around a single prior distinction.

The distinction: 1:1 versus 1:N

Verification (one-to-one): the system compares the person's biometric trait against their own stored template, to confirm they are who they say they are. The question is "is this you?" It's fingerprint clocking-in checked against your own record, facial unlock checked against your own profile.

Identification (one-to-many): the system compares the trait against a database of templates to work out who the person is. The question is "who are you, out of everyone?" It's the camera that recognises anyone who walks through the door.

The difference isn't technical — it's about the nature of the risk: identification requires a centralised biometric database covering the whole workforce and, in practice, functions as permanent scanning. That's why the entire regulatory framework treats it more severely: data protection doctrine demands justifications that verification doesn't need, and the AI Act reserves some of its toughest boxes for it — remote biometric identification has its own entry in the high-risk Annex III, and, in the workplace, the prohibition that matters is Article 5(1)(f) — inferring the emotions of a person in the workplace, except for medical or safety reasons — rather than the real-time remote biometric identification of Article 5(1)(h), which is aimed at law enforcement in publicly accessible spaces. 1:1 verification, by contrast, is expressly excluded from that Annex III entry where its sole purpose is to confirm that a person is who they claim to be.

The immediate consequence for any company: if verification solves your case, don't deploy identification. It's the first design decision, and the cheapest one to get right early.

The test the GDPR requires: was biometrics necessary?

With the "how" settled, the "whether" remains. The GDPR impact assessment contains "an assessment of the necessity and proportionality of the processing operations in relation to the purposes", a test worth running before buying the reader: is there a less invasive alternative that reasonably achieves the same purpose? For attendance control, there almost always is — a card, a PIN, an app. That doesn't rule biometrics out; it turns it into something that has to be justified in writing: why the alternatives fall short in your specific case (documented clocking-in fraud, specific security requirements), what safeguards you're adding, and why the balance comes out in favour.

That reasoning lives in the document this scenario demands almost by definition: the Data Protection Impact Assessment (DPIA). Biometric data used to uniquely identify a person and, often, systematic monitoring of the workforce: those are two criteria on the list published by the AEPD (Article 35(4)), which specifies when Article 35 GDPR requires the assessment: it considers it necessary in most cases where two or more are met, so the DPIA here is the starting point. What happens when it is missing is covered in the article on AEPD enforcement.

The rest of the file

With the distinction made and proportionality reasoned through, the file is rounded out with the familiar pieces: a reinforced legal basis under Article 9 (in the employment context — delicate ground: the GDPR warns that consent should not be a valid legal ground where there is a clear imbalance between the data subject and the controller, recital 43, so the applicable exception has to be identified carefully, not assumed); clear information for the workforce and, where one exists, its representatives — the same duty Spain's LOPDGDD sets for cameras and geolocation; technical measures matched to the data (encrypted templates, local or distributed storage rather than a central database, defined erasure periods); and a record of all of it.

One final reading rule, because this is ground where generalisations fail in both directions: "workplace biometrics is prohibited" is false; "workplace biometrics is high-risk" as a blanket statement is false too. The correct answer is more demanding: the specific use decides — 1:1 verification for clocking in, properly justified and documented, is defensible; 1:N identification demands reasons of a different order; and some uses, such as inferring emotions at work, are prohibited except for medical or safety reasons. The work of a serious company isn't memorising the slogan — it's classifying its own case, justifying it, and being able to show the justification.

Content in accordance with Articles 9 and 35 and recital 43 of Regulation (EU) 2016/679, the AEPD's Article 35.4 list, and Article 5 and Annex III of Regulation (EU) 2024/1689.

This article is for informational purposes only and does not constitute legal advice.

Frequently asked questions

Is fingerprint clocking-in high-risk under the AI Act?

Not under Annex III: its point 1(a) excludes AI systems intended for biometric verification whose sole purpose is to confirm that a person is who they claim to be. It is still processing of biometric data, to which Article 9 GDPR applies.

Is an impact assessment needed?

Most likely. Article 35 GDPR requires it where the processing is likely to result in a high risk, and the list published by the AEPD (Article 35(4)) specifies the types of processing: it considers it necessary in most cases where two or more of its criteria are met; using biometric data to uniquely identify a person is one, and systematic monitoring of the workforce may be another.

Can the company rely on the worker's consent?

With great caution: recital 43 GDPR warns that consent should not be a valid legal ground where there is a clear imbalance between the data subject and the controller. The applicable Article 9 exception has to be identified case by case.

Sources

  • Regulation (EU) 2016/679 (GDPR), Article 9(1)
  • GDPR, recital 43
  • GDPR, Article 35(7)
  • AEPD, list of processing operations requiring a DPIA (Article 35(4) GDPR; Spanish only)
  • Regulation (EU) 2024/1689, Article 5(1)(f)
  • Regulation (EU) 2024/1689, Annex III, points 1 and 4

Last substantive review: 12 September 2026

Scope: Companies considering biometric systems for their workforce. Separates verification from identification with Annex III, point 1, and Article 5(1)(f) of the AI Act, and the processing with Articles 9 and 35 GDPR, recital 43 and the AEPD list; does not develop Annex III, point 1(b) and (c).

Reviewed by Rafael Luque

Want to start from the beginning? The Route: AI governance in seven steps →

Drafted with AI support and published after human review. Editorial responsibility: Alethexis SL.

Related

  • Articles 87 to 91 LOPDGDD: AI in the Spanish workplace has its own national rules, and they do not wait for 2027
  • FRIA and DPIA are not the same thing: why one does not replace the other
  • The AEPD's Article 35(4) list: when your AI requires a DPIA even if it is not high-risk under the AI Act
  • The AEPD already supervises AI that processes personal data: oversight that doesn't wait for the AI Act

Where to go next

Next read

Articles 87 to 91 LOPDGDD: AI in the Spanish workplace has its own national rules, and they do not wait for 2027

Next step

Take the AI Act assessment (3 minutes)

Get analysis like this in your inbox

Alethexis regulatory and product news. No noise.

I agree to receive communications from Alethexis: content about AI and regulation, and product news. I can unsubscribe at any time.

Data protection — newsletter

Controller: ALETHEXIS, S.L. (NIF B88758057). Purpose: to send you the Alethexis newsletter (content about AI and regulation, and product news). Legal basis: your consent (Art. 6(1)(a) GDPR), which you can withdraw at any time. Recipients and transfers: we use providers acting on behalf of Alethexis for hosting, security and email delivery. No disclosure of your data to third parties for their own purposes is planned, except where required by law. Some providers may process data outside the EEA with the safeguards set out in the additional information. Rights of access, rectification, erasure, objection, restriction and portability: privacidad@alethexis.com. You may lodge a complaint with the Spanish Data Protection Authority (AEPD, www.aepd.es) or with the supervisory authority of your Member State (Article 77 GDPR). More information in the privacy policy.