International disputes regarding AI solutions are conflicts concerning AI systems between vendors, deployers, and users in different countries. At Musch Legal, we expect growth in AI disputes due to EU AI Act 2024/1689. Typical triggers: AI hallucinations involving harm, bias discrimination, IP claims on training data, performance issues, and confidentiality risks. The first court rulings in the EU are expected to appear starting in 2025. A proactive legal approach is essential.
What is the legal problem? (Which AI disputes arise?)
AI disputes range from hallucinations involving harmful advice, bias discrimination against protected groups, IP claims (copyright on training data), privacy infringement, performance issues, and confidentiality risks via prompt data. For enterprises: a difficult evidentiary position (black box problem) and unfamiliarity with the legal framework. The EU AI Act creates new obligations and risks for high-risk AI; AI Liability Directive (proposal COM/2022/496) eases the burden of proof for victims.
What does the law say? (Which frameworks apply to AI?)
EU AI Act (Regulation 2024/1689), entering into force on 1 August 2024, regulates AI on a risk-based basis: prohibited AI (from 2 February 2025), high-risk AI with conformity assessment (from 2 August 2026), limited risk transparency. Fines up to 35 million euros or 7% of global turnover. AI Liability Directive (proposal COM/2022/496) for civil liability: presumed causation and disclosure rights. Product Liability Directive 2024/2853 for damage caused by AI products. For copyright: DSM Directive 2019/790 text-and-data mining-exception (Article 4 with opt-out).
For IP: training data claims via Berne Convention and national laws.
AI dispute
Legal basis
Remediation
Hallucinations
Breach of contract + tort
Damage
Bias discrimination
GDPR Art 22 + EU AI Act
Fine + damages + injunction
IP training data
Copyright, DSM Directive 2019/790
License + damages + injunction
Privacy prompts
GDPR Art 6 + AI Act Art 10
Fine + damages
Performance failure
Contract + Wbb (trade secrets)
Damages + termination
AI dispute
Legal basis
Recourse
Hallucinations
Breach of contract + tort
Damage
Bias discrimination
GDPR Art 22 + EU AI Act
Fine + damages + injunction
IP training data
Copyright, DSM Directive 2019/790
License + damages + injunction
Privacy prompts
GDPR Art 6 + AI Act Art 10
Fine + damages
Performance failure
Contract + Wbb (trade secrets)
Damage + termination
What risks do companies face? (What risks in an AI dispute?)
Fines under the EU AI Act up to 7 percent of global turnover (violation of prohibited AI), 3 percent (other), 1 percent (incorrect info). Civil damages plus reputational damage. For high-risk AI: conformity assessment required; non-compliance results in loss of market access. For IP claims on training data: in the US, multiple class actions are pending against OpenAI, Microsoft, and Google. For enterprise-deployers: indirect liability via defective AI tool. Cyber policies often do not cover AI claims.
Practical example from our practice (How did we limit AI liability?)
Musch Legal advised a Dutch financial services company on an AI credit scoring tool from an American supplier. Discovered after 8 months of use: the tool exhibited systematic bias against a specific group of applicants. We pursued parallel actions: contract termination (cure period ignored by supplier), damages claim for remediation (1.2 million euros), GDPR notification to the DPA, and complaint under the EU AI Act (high-risk in effect since August 2026). Settlement: 950,000 euros plus full termination without penalty. A rapid legal response prevented both a regulatory fine and customer claims.
What can you do? (What AI strategy are you building?)
Implement AI governance before deployment: risk assessment per use case, vendor due diligence, contractual guarantees (no IP infringement, performance, bias testing). Test AI suppliers for EU AI Act compliance (high-risk if applicable). For data: opt-out of text and data mining under DSM Directive 2019/790. Build an incident response procedure for AI failures. Insure AI risk via dedicated cyber policies. Engage Musch Legal for AI governance.
Artificial intelligence and international law