AI law and works constitution law
The AI Act and German works councils: which co-determination rights apply?
The AI Act does not replace the works council’s co-determination rights under the BetrVG.

AI in recruitment and personnel management raises new questions for employers and works councils. The EU AI Regulation supplements the legal framework, but participation rights continue to arise mainly from the Works Constitution Act.
What role does the works council play when AI is introduced?
The AI Regulation regulates providers and deployers of AI systems and in part contains employment-related obligations. Co-determination at establishment level continues to be governed mainly by the Works Constitution Act. Article 2(11) of the AI Act permits more favourable provisions protecting employees and collective agreements. What matters is therefore the interplay of EU AI law, the BetrVG, data protection and the specific purpose of use.
When does section 87(1) No. 6 BetrVG apply?
Where technical devices that are objectively suitable for monitoring the behaviour or performance of employees are introduced or used, a mandatory co-determination right may exist under section 87(1) No. 6 BetrVG. An actual intention to monitor is not always required. Systems for measuring productivity, analysing communication or creating personal performance profiles must therefore be examined particularly carefully. (Section 87 (1) No. 6 of the BetrVG)
However, this does not give rise to a general right of veto against any use of generative AI. What is decisive are the functions of the specific system, how it is used and other co-determination triggers.
What information rights are added?
Under section 90 BetrVG, there are rights to planning, information and consultation, among other things, regarding work procedures and the use of AI. Section 95(2a) BetrVG clarifies that co-determination rights on selection guidelines also apply where these are created with AI support. Under section 80(3) sentence 2 BetrVG, calling in an expert may be necessary for AI; the details of the commissioning remain to be clarified. (Section 90 of the BetrVG · Section 95 (2a) of the BetrVG)
What does Art. 26 (7) AI Act change?
For high-risk AI in the workplace, Article 26(7) of the AI Act provides for informing worker representatives and affected employees. Amending Regulation (EU) 2026/1744 has, in principle, postponed the specific requirements for high-risk AI under Annex III until 2 December 2027. National information and co-determination rights and data protection obligations already apply independently of this. (Art. 26 para. 7 AI Act)
How should an AI works agreement be designed?
- Name systems, purposes, user groups and data flows specifically.
- Set out permitted and excluded monitoring purposes.
- Regulate roles, access rights, deletion periods and human control.
- Define information rights and procedures for system changes.
- Integrate a data protection impact assessment and risk classification where needed.
- Agree regular reviews and qualification measures.
For a comprehensive overview for employers, see AI in human resources. Kanzlei Osuji in Munich advises on the legal design of AI projects. (AI in HR)
Frequently asked questions
Can the works council ban any AI?
No. Co-determination depends on the relevant statutory provisions and the specific application.
Do companies have to do nothing until December 2027?
No. GDPR, the AGG, the BetrVG and the parts of the AI Regulation that already apply must still be observed.
If you have any questions, please contact Kanzlei Osuji or call +49 174 343 0142.
As of 7 November 2026. Author: Noah Osuji, lawyer. General legal information, not individual advice. AI-generated symbolic image without real participants.
AI-related change to the business: defining the conciliation committee’s remit precisely
On 31 August 2026 (9 TaBV 38/26), the Cologne Regional Labour Court dealt with a conciliation committee on a reconciliation of interests and a social plan for an AI-related change to the business. In the proceedings to appoint the committee, its remit could not, under the specific circumstances, be extended against the will of the other side to cover the introduction and use of AI systems generally, as no specific need for regulation had been shown in that respect.
The decision does not mean that AI systems are generally exempt from co-determination. Information rights, § 87 BetrVG or questions of qualification under § 97 BetrVG may require separate reviews and regulation. What matters is to define the actual conflict and the responsible body precisely. The guide to AI-related workforce reduction, social plans and conciliation committees sets the procedures side by side. For individual duties and the consequences of mistakes, the article AI use and termination adds a further perspective. (AI-based workforce reduction, social plan and Reconciliation Point)
Also: AI in the workplace: when mistakes can lead to dismissal. (AI in the workplace: when can mistakes justify dismissal?)
For advice on your situation: Legal advice on AI and IT.
