Gallery

Contacts

405 W. Greenlawn Ave Lansing, Michigan 48910

contact@techjacksolutions.com

+1-616-320-4064

Skip to content
S SEC Regulation
Regulation Daily Brief

Italy's Draft AI Decrees Would Ban Automated Employment Decisions and Require AI Training for Lawyers and Doctors

2 min read Jonesday Partial Weak S
New legal analysis published this week details Italy's draft AI implementing decrees, documents that reportedly go beyond EU AI Act baseline requirements to prohibit fully automated employment decisions and mandate AI training for licensed professionals. The decrees, which received preliminary cabinet approval on June 10, aren't yet in force, but they describe the most developed national EU AI Act implementation framework among Member States according to legal analysts.

Key Takeaways

  • Italy's Council of Ministers gave preliminary approval to draft EU AI Act implementing decrees on June 10, 2026, decrees are not yet in force
  • Draft decrees reportedly prohibit fully automated employment decisions, requiring documented human review (per DLA Piper legal analysis)
  • Licensed professionals (lawyers, doctors, engineers) would reportedly face mandatory AI training mandates under the draft decrees
  • EU AI Act Article 50 transparency obligations remain on track for August 2, 2026, Italy's national layer is a separate, later compliance problem

Warning

DRAFT, NOT YET IN FORCE. Italy's implementing decrees must still complete parliamentary committee review, State-Regions Conference review, and Official Gazette publication before they have legal effect. All Italy-specific provisions below are drawn from law firm analysis of preliminary text, not primary government sources.

Compliance Deadline

August 2, 2026
32 days remaining
EntityEU AI Office / All Member States including Italy
JurisdictionEU
PenaltyUp to 1.5% of global annual turnover (non-compliance with Article 50)

DRAFT, NOT YET IN FORCE. Italy’s Council of Ministers granted preliminary approval to two draft legislative decrees in early June aimed at implementing the EU AI Act within domestic law. The underlying approval happened June 10. What’s new this week is the legal analysis: publications from DLA Piper and Jones Day, published June 25, now detail what these decrees reportedly require, and several provisions go well beyond what the EU regulation itself mandates.

The August 2 deadline isn’t moving. EU AI Act Article 50 transparency obligations, including mandatory labeling of deepfake content and AI-generated synthetic media, take effect August 2, 2026 across all Member States. Italy’s draft decrees sit on top of that baseline, adding a national layer for companies operating in Italy specifically. Per Regulation (EU) 2024/1689, Article 57 also requires each Member State to have at least one operational national AI regulatory sandbox by that same date.

What Italy’s draft adds, all of the following requires qualification: these are preliminary decrees, sourced from law firm analysis, not primary government text.

EU AI Act Baseline vs. Italy Draft Addition (Employment Decisions)

EU AI Act (all Member States)
High-risk AI in employment classified under Annex III; human oversight required under Article 14; technical documentation and logging required
Italy Draft Decree Addition
Reportedly makes 'meaningful and documented' human review standard explicit for hiring, role changes, discipline, and termination, per DLA Piper analysis of draft text

According to legal analysis of the draft decrees, Italy would require documented human review for AI-assisted employment decisions including recruitment, role changes, discipline, and termination. The EU AI Act’s Annex III already classifies employment AI as high-risk; Italy’s draft reportedly makes the “meaningful and documented” standard explicit. The draft decrees reportedly also amend Italy’s Industrial Property Code to extend trade secret protections to AI training data and algorithms, legal analysis cites Article 98 of the Code specifically, though that provision hasn’t been confirmed against primary text. And licensed professionals, lawyers, doctors, and engineers, would reportedly be required to complete AI training programs covering technical limitations, legal frameworks, and professional responsibilities, with professional bodies given six months to implement and forensic fee schedules given 12 months.

Legal analysts characterize Italy’s draft decrees as potentially the most developed national EU AI Act implementation framework among Member States. That claim hasn’t been confirmed against other Member States’ implementation activity, treat it as analytical framing, not established fact.

Italy Operations: Draft Compliance Preparation Steps

  • Complete EU AI Act Article 50 gap assessment (applies August 2 regardless of Italy decrees)
  • Inventory AI systems used in employment decisions affecting Italian employees
  • Assess whether current HR AI audit trails meet 'meaningful and documented' standard
  • Identify licensed professionals (lawyers, doctors, engineers) in Italian operations who may need AI training
  • Monitor Official Gazette for final decree publication

The catch is timing. These decrees are still draft. They must undergo parliamentary committee review, State-Regions Conference review, and formal publication in the Official Gazette before entering into force. Companies with Italian operations can’t build compliance programs against text that isn’t final. What they can do: use this analysis to scope what’s likely coming, particularly the employment decision documentation requirements, which align with what the EU AI Act already requires for high-risk systems and are unlikely to be removed in review.

Don’t expect the August 2 deadline to wait. Whatever Italy’s final implementing decrees say, Article 50 transparency obligations arrive in five weeks. Compliance teams with EU operations should have already resolved their Article 50 gap assessment. Italy’s national layer is the next problem, but it isn’t August’s problem.

View Source
More Regulation intelligence
View all Regulation

Stay ahead on Regulation

Get verified AI intelligence delivered daily. No hype, no speculation, just what matters.

Explore the AI News Hub