Prepare for the AI Act
without waiting for December 2027.
For CHROs, legal teams and leadership wondering what the European regulation on AI (AI Act) actually changes. One obligation is already in force; another is coming for AI tools used to recruit, assess or assign tasks. There is still time to get organized, and we help you make good use of it.
“Does this apply to us?”
The legal team raises the question in a meeting. Your recruiting software sorts applications with AI. Your training tool recommends learning paths. Managers prepare annual reviews with an assistant. No one has taken stock.
AI in HR is not banned, it is regulated. The employer that uses the tool is accountable for its compliance, even if it didn't build it. There is still time to get organized, as long as you start with that inventory.
What applies, and from when.
- February 2, 2025
Train staff in AI
Employers must ensure their staff have sufficient literacy in the AI they use (Article 4). Already in force.
- February 2, 2025
Prohibited practices
Among other things, using AI to infer people's emotions at work is prohibited, except for medical or safety reasons.
- December 2, 2027
High-risk HR tools
Recruiting, assessing, promoting, dismissing, assigning tasks, monitoring performance: AI tools used for these decisions will have to meet strict rules.
For high-risk HR tools.
- Inform employee representatives and the people concerned before the tool is put into service. On this point: talking about AI with the works council.
- Assign human oversight to trained people who have the authority to intervene.
- Be able to explain a decision to the person concerned when they ask, in recruiting as well as on the job.
A head start on the deadline.
Since AI changes tasks far more than whole jobs, looking at tasks is how you see where it steps in. Our method, step by step.
Your AI training plan
The skills to develop, job by job. Take recruiting: checking a sort done by AI, spotting a bias, explaining a decision. Enough to meet Article 4, and to inform your workforce planning.
The map of AI across the company's jobs
Where it steps in, where human oversight is required, what stays human. If your AI policy requires systematic oversight, you immediately see what that changes.
A dossier for your works council
A job-by-job report, to share before any deployment.
What HR leaders and legal teams ask.
What are the penalties?
For employers that use a high-risk tool without meeting their obligations: up to €15M or 3% of worldwide turnover. The ceiling of €35M or 7% applies to prohibited practices.
Is Skillberg itself a high-risk tool?
The tool analyzes jobs, never people: it rates no one, sorts no applications and makes no individual assignment decisions. All the details in our commitments.
What should we do right now?
Train your teams in AI, since it is already mandatory, identify where AI steps in across your jobs and HR tools, then prepare to inform the works council. The AI Act and works council kit guides you step by step.
Get ahead of December 2027.
In 30 minutes, you tell us where you stand, we show you on a fictional company how the tool prepares an AI training plan, and together we sketch the first steps for your company.
30 minutes on video with Skillberg, no commitment. No personal data required.
- Regulation (EU) 2024/1689 on artificial intelligence, as amended by the 2026 “Digital Omnibus” regulation: Articles 4, 5, 26, 86 and 99, and Annex III