False Positive Safety Block / Error with Custom AI Mascot Reference Image in Veo

Hi everyone,

I am encountering a recurring issue when generating videos with Veo in Google AI Studio using reference images, and I suspect a false positive safety or filter trigger is blocking the request.

Setup & Workflow:

  • Model: Veo 3.1 fast in Google AI Studio (using a paid API key with topped-up credits)

  • Goal: Generating a video combining a reference image of a historic building facade with a reference image of our brand mascot (a fox).

The Issue: Last week, this exact setup and prompt worked as expected. However, the generation now fails whenever I include the reference image of our fox mascot.

Troubleshooting Conducted:

  1. Text prompt only: Works fine.

  2. Building facade reference image only: Works fine.

  3. Fox mascot reference image included: Fails with an error message.

Important Context: The fox mascot is an original, AI-generated character created specifically for our brand. It does not infringe on any third-party copyright, nor does it contain any sensitive or restricted content.

Error Details: Failed to generate a video. Your prompt was blocked due to safety reasons. or Failed to generate video: Your use case is currently not supported. Please refer to Gemini API documentation for current model offering.

Has anyone experienced similar false positives recently? Is there a known workaround, or could a team member take a look at the logs to help resolve this?

Thanks in advance!

Hello @MediaAI_Creator ,

Thanks for the detailed breakdown, that isolation is helpful.

Two things would let us pin this down.
First, how is each image passed in your request — is the building the starting image and the fox a reference image, or are both reference images?

Second, which of the two errors do you get when the fox image is included, since those two messages come from different checks.

If you can also share the mascot image with us directly, we can test it against the same setup.

Thank you!