Gemini AI may refuse or restrict a request when it believes the request could create safety, privacy, legal, or quality concerns. In many cases, the limitation is intentional and is tied to policies that are meant to reduce harm, especially for requests involving dangerous instructions, illegal activity, explicit sexual content, personal data, medical or legal advice presented as definitive, or attempts to manipulate the system.
Sometimes a response is also limited because the request is too broad, unclear, or missing important context, so Gemini AI may choose a cautious answer instead of guessing. It may also restrict content if the prompt asks it to identify a private person, reveal sensitive information, or produce content that could be misleading if taken as fact.
If this happened on a specific page or product, it can help to reread the prompt for any sensitive wording and check the current web page for any guidance, notices, or contact information. Often, a small rewrite that makes the request safer and more specific will produce a better result.