EU just dropped €47M in fines on hiring AI, credit scoring, and emotion recognition. Zero days after enforcement started.

The EU AI Office issued its first fines 17 days after enforcement went live.

Three companies. Three violation types. €47 million total.

€18M to an HR tech company for deploying hiring AI without conformity assessment documentation or human oversight controls.

€14M to a credit scoring provider for Annex III documentation failures.

€15M to a retail chain running real-time emotion recognition across four EU member states. A prohibited practice under Article 5.

Here is what matters: these cases were pre-built before August 2. The AI Office had a queue ready on day one.

No ramp-up period. No informal warnings. No grace period.

If your company uses AI in hiring, credit decisions, or anything that infers emotional states from facial expressions or voice data — you are already in scope.

The hiring fine alone should alarm every HR tech vendor. The system had no documented conformity assessment. No designated human reviewer. No override protocol. No logging of how often human judgment was applied.

That is the baseline. Not the exception.

CNIL sent information requests to 14 French financial institutions the same week. The enforcement posture is simultaneous at EU and national levels.

Audit every AI system touching EU citizens today. If you cannot produce Annex IV documentation for high-risk systems, you are already behind. The next wave of fines is in progress.

SOURCE: https://www.aipolicydesk.com/blog/eu-ai-act-first-fines-47-million-hiring-emotion-recognition-2026
VERIFIED: EU AI Office enforcement announcement, European Commission Digital Strategy, WSJ enforcement phase coverage, CNIL credit scoring investigation
SIGNAL: The EU AI Act moved from paper to penalties in 17 days. Every enterprise deploying AI in EU markets now faces the same enforcement speed. This is the new compliance baseline.