On 2 June 2026, in its judgment 9C_235/2026, the Swiss Federal Supreme Court dismissed a health insurance dispute complaint that had been “obviously drafted with so-called artificial intelligence (AI)” by an unrepresented individual. The court described the pleading as containing a “blurry accusation” and “interchangeable submissions without substance”, and ultimately did not even enter on the complaint.

This case is a clear signal: generative AI can produce text, but it cannot develop legal strategy. Anyone who relies too heavily on AI without any legal expert review risks not only losing the case, but also bearing court costs and worsening their legal position.

What the Federal Supreme Court concretely decided on the AI complaint

The complainant challenged a decision of the Lucerne Cantonal Court in a dispute over outstanding health insurance premiums. He submitted two pleadings that were “obviously drafted with so-called artificial intelligence (AI)”, as the Federal Supreme Court noted.

The complaint contained general objections (e.g. “contradictory file situation”, “lack of transparent presentation”), but no concrete references to the reasoning of the lower court. The court found:

• a “blurry accusation” that did not “in any way” relate to the outstanding premiums;
• “interchangeable submissions without substance”;
• no legally sufficient statement of claim under Art. 42 BGG.

The Federal Supreme Court therefore, in a simplified procedure, did not enter on the complaint. In this case, no court costs were imposed; however, had the court seriously dealt with the “arguments”, the advance on costs alone would have been CHF 500 – cost can rise up to CHF 5’000 even with low amounts in dispute.

What companies and private individuals must learn about using AI in legal matters

This judgment is not a blanket rejection of AI, but a clear warning: AI is a tool, not a substitute for legal responsibility.

• For companies:
  If you use AI to draft contracts, compliance statements, or data protection documents, every AI-generated text passage must be reviewed, adapted, and fact-specifically substantiated by a qualified professional (e.g. data protection officer, legal department). The risk of “generic statements” and “submissions without substance” is just as high in practice as in court proceedings.
• For private individuals:
  Using AI for initial orientation is useful, but a complaint, application, or lawsuit that is purely AI-generated and submitted without legal advice is likely to end in a blunder like in case 9C_235/2026, with possible loss, additional costs, and the lower court decision becoming legally binding .

Our recommendation: AI as an assistant, lawyer as decision-maker

From our practice as lawyers specializing in data protection and digital matters, we recommend:

• Use AI for draft text, structural help, and wording variants.
• Have every legally relevant passage reviewed, adapted, and fact-specifically supported by a specialized legal professional.
• Document this review internally, especially under the GDPR and the new AI Act, to demonstrate human responsibility and control.

The Federal Supreme Court’s ruling is unequivocal: anyone who relies too heavily on generative AI and skips legal expert review will end up producing only “blurry accusations” and “submissions without substance” – both before courts and in business life.

If you need support on the legally compliant use of AI in your company – for example regarding data protection, liability, and document review – please contact us. We will assist you from the AI policy to court-ready and lawyer-reviewed contract and complaint practice.