Run the free classifier
7 yes/no questions about the specific workforce-AI system you are assessing. Instant indicative band, the Article 6(3) exemption test, and the obligation cascade — free, no email needed to see your result. An optional email step inside the tool sends you the full classification trace as a PDF.
What this classifier does
The classifier walks the same decision tree counsel walks when scoping a workforce-AI system under the EU AI Act (Regulation 2024/1689). Question 1 confirms the Annex III point 4 trigger — employment or worker-management use. Questions 2–5 test profiling, decision influence, sensitive signals (biometric, emotion inference, continuous screenshot or keystroke capture), and EU nexus. Questions 6–7 run the Article 6(3) exemption test: narrow procedural task, or improving the result of prior human activity.
The output is one of three bands — Not high-risk, Likely high-risk (Annex III point 4), or Definitely high-risk + obligations — each mapped to the articles that drive your next 12 months of work. It is an awareness and planning aid, not a conformity assessment: the band tells you whether to start the Article 9 and Article 27 workstreams now, not whether you have passed them.
Unlike the all-sector classifiers that rank for this query, this tool only does the workforce category — which means the questions can be specific enough to be useful: it asks about appraisal influence and keystroke capture, not abstract “intended purpose” categories.
How Article 6 classification works
Article 6(1) and 6(2) set the two routes into high-risk: a system that is a safety component of regulated products (6(1)), or a system whose intended purpose falls in an Annex III category (6(2)). Workforce AI enters through the second route.
Annex III point 4 is the workforce category. It covers AI intended for recruitment or selection of people, for decisions affecting terms of work, promotion, or termination, for task allocation based on behaviour or traits, and for monitoring and evaluating the performance of workers. Most productivity-monitoring, screening, and workforce-analytics tools sit squarely in this language.
Article 6(3) is the exemption test. A system in an Annex III category escapes high-risk classification only if it performs a narrow procedural task, improves the result of a previously completed human activity, detects decision patterns without replacing human assessment, or performs a preparatory task — and the exemption is documented before deployment. The carve-outs are construed narrowly.
The practical trap: profiling defeats the exemptions. A system that builds behavioural or performance profiles of workers cannot ride the Article 6(3) carve-outs, which is why most monitoring tools that score individuals come out Likely or Definitely high-risk on this tree.
What a high-risk result means
A high-risk band is not a prohibition — it is a workplan. These are the obligations the classifier maps your band to:
| Obligation | What it requires |
|---|---|
| Article 9 — risk management | A documented, iterative risk-management system across the AI lifecycle |
| Article 10 — data governance | Training, validation, and testing data quality and bias controls |
| Article 11 — technical documentation | Annex IV technical file before placing on market or into service |
| Article 13 — transparency | Instructions and information sufficient for deployers to use the system compliantly |
| Article 14 — human oversight | A named human who can interpret, override, or halt the system’s output |
| Article 26 — deployer obligations | Use per instructions, assign oversight, monitor operation, inform workers |
| Article 27 — FRIA | Fundamental rights impact assessment for certain deployers before first use |
Penalty line for the file — non-compliance with most high-risk obligations carries penalties under Article 99 of up to EUR 15M or 3% of worldwide annual turnover, whichever is higher. Run the numbers alongside your DPDP/GDPR exposure with the free GDPR + DPDP Fine Exposure Estimator. Verify with counsel.
For India IT exporters
The EU AI Act applies by territorial scope, not incorporation: if your system’s output is used in the EU, or you place the system on the EU market, the Act applies regardless of where you are established. For an India IT-services exporter running workforce AI — a monitoring stack over teams that serve EU clients, a scheduling engine whose allocations land in an EU delivery centre — that is the live question, and it is Question 5 in the classifier.
Two India-specific notes the generic classifiers skip. First, Article 22: a provider established outside the EU must appoint an authorised representative in the EU before making a high-risk system available there. Second, the EU AI Act lands on top of an existing DPDP Act 2023 workload — the same monitoring system that is high-risk under Annex III point 4 usually also processes employee personal data that needs DPDP notice and purpose answers. The EU AI Act time-tracking compliance guide walks the overlap, and the workforce monitoring compliance checklist sequences the work.
Three minutes, no signup — classify the system you are thinking about right now: open the EU AI Act Article 6 Classifier →
Frequently asked questions
Is my workforce AI high-risk under the EU AI Act?
A workforce-AI system used to recruit, evaluate, allocate work to, monitor, or make decisions affecting workers is classified as high-risk under EU AI Act Article 6(1) read with Annex III point 4 (employment, workers management, and access to self-employment). The Article 6(3) carve-outs for narrow procedural tasks, improving prior human activity, or preparatory tasks rarely apply once the system profiles individuals or influences an HR decision. High-risk obligations apply from 2 August 2026. The free classifier returns an indicative band only — verify with counsel.
What is Annex III point 4 of the EU AI Act?
Annex III point 4 lists AI systems intended to be used for the recruitment or selection of people, for decisions affecting terms of work, promotion, or termination, for task allocation based on behaviour or traits, and for monitoring and evaluating performance of workers. Systems in this category are high-risk under Article 6(2) unless an Article 6(3) exemption is properly documented.
When do EU AI Act high-risk obligations start?
High-risk system obligations under the EU AI Act (Regulation 2024/1689) apply from 2 August 2026. Prohibited-practice provisions applied earlier. Exact timing for specific obligations is subject to guidance — verify with counsel.
Does the EU AI Act apply to an India IT services exporter?
Yes, where the AI system's output is used in the EU or the provider places the system on the EU market, the EU AI Act applies regardless of where the provider is established. India IT services exporters building or operating workforce AI for EU customers may also need an Article 22 authorised representative in the EU. Verify with counsel.
Classified high-risk? Score your vendor next.
Run your monitoring or workforce-AI vendor through the 14-question EU AI Act Vendor Scorecard — live verdict band, Article 5 halt conditions, free to score. Or talk through your band with the gStride team.
Disclaimer: This page and the linked classifier are general information and an awareness aid, not legal, compliance, or conformity-assessment advice. The classifier returns a qualitative, indicative band from seven yes/no inputs; it is not a formal Article 6 determination. The EU AI Act (Regulation 2024/1689) is in force, high-risk obligations apply from 2 August 2026, and Annex III scope and Article 6(3) carve-outs remain subject to guidance. Verify classification, obligations, and penalties with qualified counsel before acting. gStride is not certified or approved by any EU authority.