Employers do not need the consent of the works council to allow their employees to (voluntarily) use ChatGPT at work if they use their private accounts. This was ruled by the Hamburg Labour Court on 16 January 2024 (24 BVGa 1/24).
In its request, the works council claimed that its co-determination rights had been grossly violated by authorising the use of ChatGPT. The employer had previously authorised its employees to use ChatGPT as a tool, provided that they used their private accounts. The Hamburg Labour Court rejected the application – it saw no violation of the co-determination rights pursuant to Section 87 Works Constitution Act, as the AI application had been used by the employees voluntarily and exclusively via the browser. Therefore, the Labour Court did not see any risk of surveillance pressure arising, as the employer had no access to the data collected by the AI operator. The Hamburg Labour Court was the first labour court to take a comprehensive stance on the issue of co-determination in the use of artificial intelligence.
The use of artificial intelligence in the workplace is a new subject area in which there is still little case law. However, it can be inferred from the judgement that the works council has no right of co-determination when employees voluntarily use AI tools via their private accounts. This should be accepted, and the ruling provides an opportunity to take a closer look at the rights of the works council in relation to AI applications. With the Works Council Modernisation Act, the legislator extended the Works Constitution Act in 2021 with regard to AI, so that the following (non-exhaustive) picture currently emerges:
Labour law practice in relation to AI is currently in a state of flux. It remains to be seen to what extent the use of artificial intelligence will be accepted in companies. It cannot be ruled out that some works councils are reluctant to introduce AI because the associated dangers and risks cannot yet be conclusively assessed. In some companies, however, the first agreements on AI systems are already being negotiated between employers and works councils. From the employer's point of view, however, the latter is not necessary, at least if employees voluntarily use AI systems via private accounts.
However, in this scenario, it cannot be ruled out that confidential information and data will become public (e.g. because the AI learns with the data from the private user account and passes it on to another external user request). For this reason, employers are advised to generally only allow employees to use AI and IT systems and accounts provided by the employer, as this is the only way to prevent confidential information from being leaked, if at all.