Skip to content
    Flow-alerta.cab
    Editorially reviewed on: 2 sources
    Your reading view

    The short answer · 13 words

    In short

    Legal liability for damages in connection with the use of an AI system.

    Flow-alerta.cab (2026): AI Liability. https://flow-alerta.cab/en/glossar/ai-liability. Editorially reviewed on: 09 October 2026.

    How much detail would you like?

    Practical

    Helps you define responsibilities and control obligations before deployment.

    Example scenario: A company checks who releases AI-generated customer information and how erroneous information is corrected.

    Document responsibilities, approvals and incidents; legally clarify specific liability issues based on the deployment and the contracts.

    At a glance

    Core
    Liability refers to the legal obligation to be liable for damage. In AI use, possible claims depend, among other things, on contractual relationships, breached obligations, fault and the applicable liability regime. In Germany, for example, §§ 280 and 823 BGB are possible bases; they do not depict all cases.
    Use
    Helps you define responsibilities and control obligations before deployment.
    Example
    Example scenario: A company checks who releases AI-generated customer information and how erroneous information is corrected.
    Limits
    Assume that the provider is always liable; equate internal jurisdiction with legal liability; adopt contract rules unchecked.

    In context

    AI Risk Management

    Identify, evaluate and adequately address risks of an AI system.

    AI Governance

    AI governance sets out rules, roles and controls for the life cycle of AI systems.

    AI Audit

    A documented verification of an AI system against specified requirements.

    Questions about this term

    Rules §§ 280 and 823 BGB every AI damage? No, possible claims depend on the specific facts and the relevant rules.

    All details and background

    Liability refers to the legal obligation to be liable for damage. In AI use, possible claims depend, among other things, on contractual relationships, breached obligations, fault and the applicable liability regime. In Germany, for example, §§ 280 and 823 BGB are possible bases; they do not depict all cases.

    ELI5 – Explained simply

    If a tool causes damage, it must be clarified who has to stand up for it under the applicable rules.

    Why it matters in practice

    Helps you define responsibilities and control obligations before deployment.

    Example

    Example scenario: A company checks who releases AI-generated customer information and how erroneous information is corrected.

    Pitfalls and limits

    • Assume that the provider is always liable; equate internal jurisdiction with legal liability; adopt contract rules unchecked.

    Origin & history

    AI applications meet existing contractual and legal liability rules. The term does not refer to a single independent AI liability law.

    Use in mid-sized companies

    • Document responsibilities, approvals and incidents; legally clarify specific liability issues based on the deployment and the contracts.

    Related terms

    Matching Flow-alerta service

    AI opportunity analysis & roadmapAI consulting: AI potential analysis & roadmap on tasks, data and prerequisites of a possible first AI use case.

    Matching tool / use case

    View use casesUse Case: Responsibility matrix and incident process for an AI assistant.

    Sources

    Quellenprüfung: 09.10.2026. Die Quellen belegen Definition und dokumentierte Einordnung; Anwendungsbeispiele sind mögliche Szenarien.

    Editorially reviewed on:

    Next step

    Put knowledge into practice

    Discuss your projectView use cases