Trust & transparency
How reliable is the AI?
Where you can rely on the AI, where caution is advised, and how to verify results.
Updated 4 October 2026
The AI is a powerful tool for rapid understanding – but it is not infallible. This page helps you to correctly assess the results.
What you can rely on well
- An initial orientation: What is it about, which points are relevant?
- The translation of legalese into everyday language.
- The structuring of your case and the identification of open points (gathering the facts of the case).
Where caution is advised
- Hallucinations: AI can generate factually incorrect statements that sound plausible.
- Knowledge cutoff date: Current case law or legislative changes may be missing.
- Language & Jurisdiction: Most advanced in German/English and for German law; caution is advised internationally (languages and countries).
- Bias: Models can exhibit biases.
How to verify results
- Always check critical deadlines and figures against the original document.
- For important statements, optionally use the add sources function (primary source search).
- Provide more context and repeat the action if results are thin (more context, better answer).
- For large amounts or far-reaching steps: Involve a law firm (when to consult a lawyer).
Found an error?
Report incorrect or outdated outputs using the 👍/👎 feedback on the respective result. This helps to improve prompts and templates.
Related topics: limits of the AI · what is the AI · the AI works visibly · when to consult a lawyer · more context, better answer · evidence verification
Was this article helpful?
