Responsible AI · EU AI Act
Built for the AI Act era.
The EU AI Act treats AI in hiring as high-risk, and bans the riskiest practices outright. Spoon Hire was designed from day one around the principles behind it: no banned emotion recognition, skills-first and anonymized, voice not video, and a human who makes every decision.
What the Act says about hiring
AI used to filter applications or evaluate candidates is high-risk (Annex III), carrying obligations on transparency, human oversight, accuracy and record-keeping that phase in from around 2026.
What it bans
Emotion recognition in the workplace and biometric categorisation are prohibited. Many video interview tools infer confidence or affect from a face or voice — Spoon does not.
By design
Six choices that keep hiring fair.
No emotion recognition
The Act bans inferring emotion in recruitment. We never do it — interviews are scored on the substance of your answers, not your mood or tone.
Voice, not video
No webcam, no facial analysis. Removing appearance from the room removes a whole class of bias the Act is designed to prevent.
Anonymized by default
Name, photo, school and network are hidden during early screening, so the first judgment is about the work.
A human decides
AI assists; people decide. A human recruiter invites you, reviews your interview, and makes every shortlisting and hiring call.
Deterministic tests
Skills and language tests are rule-scored, not AI — explainable and identical for everyone.
Manipulation safeguards
We detect and flag attempts to game the AI interviewer, and we keep a reviewable record of every interview.
The human in the loop is the recruiter.
Spoon’s AI is decision-support, never the decision-maker. Every step that affects a candidate is taken by a person:
- 1
A human invites
A recruiter reviews the anonymized, skills-ranked pool and chooses who to invite to interview.
- 2
AI interviews
Every candidate gets the same structured AI voice interview, scored on substance.
- 3
A human decides
The recruiter reviews each interview and makes every shortlist, connect and hire decision.
Because a person makes every consequential decision, no candidate is subject to a solely-automated rejection. After an interview, candidates can also ask the hiring team to take another look.
For employers (deployers)
If you use Spoon to hire in the EU, you have your own duties as a deployer. We help you meet them:
- Instructions for use and a record of how the AI is overseen
- A fundamental-rights impact assessment (FRIA) checklist
- A ready candidate-notice you can publish
- Exportable interview records and a hiring-decision trail
Ask us via support for the current documentation pack.
Data & privacy
Contact details stay private until a recruiter chooses to connect, and we minimise what we collect. See our Privacy Policy, Candidate privacy, Responsible AI and DPA.
Questions
The specifics.
Yes. AI used to filter applications or evaluate candidates is classified as 'high-risk' under Annex III, with obligations phasing in from around August 2026. Some practices — notably emotion recognition in the workplace — are banned outright from February 2025.
This page explains how Spoon Hire is designed and is provided for information only — it is not legal advice. Obligations depend on your role and use; consult a qualified adviser for your situation.
Hire with AI you can stand behind.
Skills-first, human-overseen, and built for the rules ahead.