Teti is providing incorrect or misleading responses. What's going on?

Teti is an AI, and in an attempt to be a helpful assistant it can occasionally produce responses that are incorrect or misleading.

This is known as "hallucinating" information, and it's a byproduct of some of the current limitations of frontier Generative AI models. For example, in some subject areas Teti might not have been trained on the most up-to-date information and may get confused when asked about current events. Another example is that Teti can display quotes that may look authoritative or sound convincing, but are not grounded in fact. In other words, Teti can write things that might look correct but are actually mistaken.

You should not rely on Teti as a singular source of truth, and you should carefully scrutinize any high-stakes advice it gives you — especially on medical, legal, or financial matters.

Working with web search results

When Teti uses web search results, review the sources it cites. Original websites may contain important context or details not included in Teti's synthesis. The quality of Teti's responses also depends on the underlying sources it references, so checking the original content helps you identify any information that might have been misinterpreted without the full context.

Working with your documents

When Teti answers questions about files you've uploaded, it's a good idea to double-check important facts against the original document, particularly numbers, dates, and quotes.

How to help us improve

• Use the reactions under a message to let us know when a particular response was unhelpful or incorrect.
• Rephrasing your question or adding more context often helps Teti produce a more accurate answer.
• You can also write to us at [email protected] with your thoughts or suggestions.