Legal

AI Disclaimer

Last updated: [PLACEHOLDER — date of legal review]

Draft — pending legal review. This page was drafted as a complete, structurally-standard policy document, but fields marked [PLACEHOLDER] below are not yet filled in with real, company-specific facts (legal entity name, registered address, governing-law jurisdiction, data protection officer contact, exact retention periods). Do not treat this page as legally binding, and do not publish it as final, until a qualified lawyer has reviewed it and every placeholder has been replaced with an accurate, verified value.

Not legal advice

The Juricare AI “AI Legal Assistant” and any AI-assisted document analysis or drafting features (“AI Features”) generate responses using a third-party large language model. AI Feature output is not legal advice, is not a substitute for a licensed attorney’s judgment, and must be independently verified before you rely on it for any legal, procedural, or filing purpose. No attorney-client relationship is formed by using the AI Features.

How this actually works, technically

This section describes the real, current behavior, not a generic AI disclaimer:

  • The primary AI provider is Google Gemini, used only when the platform operator has configured a real GEMINI_API_KEY.
  • If no AI provider is configured, the AI Features return a clearly-labeled placeholder response — the platform is built to never silently fabricate an AI-generated answer when no real model is actually connected. Document analysis results likewise carry an explicit isPlaceholder flag when this is the case, surfaced honestly in the product UI rather than hidden.
  • Your prompts (and, for document analysis, relevant document content) are sent to the configured third-party AI provider to generate a response — see the Privacy Policy for what data that involves.

Known limitations

  • Like any large language model, responses can be factually wrong, outdated, or confidently state something incorrect (“hallucination”) — this is a known, unresolved characteristic of the underlying technology, not specific to a bug in this platform.
  • The model may not reflect the most recent changes in law, regulation, or court procedure. It has no awareness of facts specific to your matter beyond what you explicitly provide in your prompt.
  • Document analysis (OCR/text extraction/summarization) can misread poor-quality scans, non-standard formats, or handwriting, and can miss or misstate content.

Your responsibility

Any advocate or firm using AI Feature output in client-facing work is responsible for independently verifying it against primary sources (statutes, case law, court rules) before relying on it, exactly as you would verify any research assistant’s work product. [PLACEHOLDER — jurisdiction-specific professional-conduct language, if your bar/regulatory body has specific guidance on AI-assisted legal work, e.g. disclosure obligations to clients or courts].