The use of artificial intelligence (AI) offers companies enormous advantages: Processes become more efficient, decisions are optimized based on data, and creative solutions emerge more quickly. But where there are opportunities, there are also risks – especially if the AI makes mistakes. Who is liable in the event of incorrect decisions, data protection breaches or financial losses? This article gives you an overview of the legal liability issues when using AI and shows you what to look out for.
Liability for AI errors: who is responsible?
A key question when using AI is: who is liable if the AI makes a mistake? Unlike a human being, who can be held personally liable for their actions, an AI is not “liable”. This means that responsibility must be placed elsewhere.
1. responsibility of the company
As the operator of the AI, the company has a central responsibility. This includes:
- Contractual liability: If the AI makes a mistake that affects contractually owed services, the company using the AI is liable. An example: An automated chatbot provides incorrect legal information that leads to a financial loss.
- Tortious liability: If a third party is harmed by the use of AI, the company can be held liable for the damage caused. This could be the case, for example, if an AI-supported health diagnosis leads to incorrect treatment.
2. manufacturer’s liability
The manufacturer of the AI may also be liable, particularly if damage is caused by faulty software. This is governed by product liability. An example: An AI for production optimization causes machine downtime due to a software error. In this case, the manufacturer of the AI may be held liable under certain circumstances.
3. employees and users
Employees who operate the AI are generally not personally liable unless they act with gross negligence or intent. However, companies should ensure that employees are adequately trained in the use of AI.
What does the legal framework say?
The EU AI Act
The recently adopted EU AI Act regulates the use of AI in Europe and prescribes strict requirements for high-risk AI systems. Companies that use and/or develop such systems must, among other things:
- Carry out risk assessments,
- Comply with transparency requirements and
- check the systems regularly.
Violations of these regulations can lead to considerable fines. A practical example: AI-supported application software unknowingly discriminates against certain groups of people – the company is liable for non-compliance with the requirements.
Data protection and GDPR
The use of AI must be compatible with the General Data Protection Regulation (GDPR). The handling of personal data is particularly sensitive:
- If an AI processes incorrect data that leads to discrimination, the company is liable.
- Inadequate data security measures can also lead to high penalties.
Practical examples of liability risks
- Incorrect price calculation by AI: An online store uses an AI that automatically calculates prices. Due to an algorithm error, products are offered at loss-making prices. The company is liable.
- Discrimination in recruiting: An AI filters applicant profiles according to criteria that have an indirect discriminatory effect. The company can be held liable here, even if the discrimination was unintentional.
- Wrong decisions in production: An AI decides autonomously on production processes and leads to massive delays due to incorrect data analysis. The manufacturer of the AI could be held responsible.
Legal security in the use of AI
The use of AI systems opens up great opportunities for companies, but also harbors legal risks. Liability issues in particular should be clarified at an early stage and safeguarded through precise contracts, training and compliance with regulatory requirements.
Our focus on the topic of “AI in the company”
The use of AI opens up exciting opportunities for companies, but also raises many legal questions. Our topic page “AI in the company” provides you with comprehensive information and practical tips on how to use AI safely and effectively.
