Artificial intelligence analyses markets, prepares decisions and recommends acquisition targets. However, if the system makes a mistake and the company suffers significant damage as a result, a crucial question arises: who is liable? The answer is legally clear-cut. Liability rests with the board of directors or the managing director, not with the machine.
The business judgement rule only provides protection where an independent assessment has been carried out
Under German company law, the so-called ‘business judgement rule’ applies. It states that members of the executive board are not personally liable if they have taken a business decision in good faith, on the basis of adequate information and in the best interests of the company. It sounds simple, but it isn’t once AI comes into the picture.
The problem is that AI possesses neither human judgement nor business experience or intuition. Yet it is precisely these elements that form the basis of the business judgement rule. Anyone who simply adopts a decision based on AI output without critically scrutinising it and drawing on their own experience cannot invoke the protection afforded by this rule. The liability risk then remains in full.
AI is not an expert opinion
If board members or managing directors are unable to form their own judgement on specific technical matters, they may rely on external expert opinions. However, this presupposes that the experts are independent, have not been given any outcome-oriented instructions, and are aware of all relevant information. As things stand, AI cannot meet these requirements. It is prone to errors, uses sources that are sometimes non-transparent, and lacks its own professional judgement. An erroneous AI result therefore does not relieve the company’s management of legal liability.
Documentation is crucial
Anyone who incorporates AI into decision-making processes must document this carefully. In the event of a legal dispute, it is essential that it is clear on what information a decision was based. The following points should be recorded:
- Name of the person who used the AI and the date of use
- AI tool used, including version and date
- Prompts used and documents uploaded
- Purpose of use and desired outcome
- How the AI results were checked for timeliness, sources and context
- Compliance with internal approval processes
As many AI tools only store results temporarily, it is advisable to save prompts and responses as PDF files.
The EU AI Act introduces additional obligations
The EU AI Act introduces further requirements for so-called high-risk AI systems: a continuous risk management system, requirements regarding data quality and human oversight, as well as documentation obligations. The so-called duty of legality obliges board members to ensure compliance with these requirements. Anyone who ignores this is in breach of their duties, with corresponding liability consequences. AI is therefore not just a technological issue, but also a compliance issue.
What this means for managing directors of limited liability companies (GmbHs)
These legal principles do not apply solely to public limited companies. Managing directors of limited liability companies (GmbHs) are also obliged to exercise due care in their decision-making. Anyone who adopts AI results without due consideration and thereby breaches their duty of care as a business owner is personally liable. Neither the IT department nor the legal department can deal with this matter on the management body’s behalf. The responsibility lies with the management body itself. Anyone who is unsure how to ensure that the use of AI within their own company is carried out in a way that minimises liability should seek legal advice at an early stage.
Source: Haufe: Law
This post and the accompanying image were created with the help of AI and have been editorially reviewed.