Why does Iris sometimes give wrong or misleading answers?

Iris tries hard to be useful, but it can sometimes produce answers that are incorrect, outdated, or stated with more confidence than they deserve.

This is often called hallucination, and it comes from the underlying limitations of the AI models Iris runs on. A few common ways it shows up:

• The model's training has a cutoff, so it can be confused or simply wrong about very recent events.
• It can generate quotes, statistics, or citations that sound authoritative but are not actually grounded in a real source.
• On unfamiliar or niche topics, it can fill gaps with something plausible-sounding rather than admitting it does not know.

Because of this, treat Iris as one input among several rather than a single source of truth. Double-check anything you plan to act on, especially for medical, legal, financial, or other high-stakes decisions.

When Iris searches the web or pulls from your own documents, look at the sources it cites rather than only the summary. The original page or file may carry context that did not make it into the answer, and the answer is only as reliable as what it found.

If you are working with information that matters, verify it independently before relying on it.