Google Generative AI Conversation Gemini Pro Safety Settings

Has anyone figured out a way to reduce the safety settings calling the Gemini Pro API?

By default, safety settings block content with medium and/or high probability of being unsafe content across all dimensions

I’m getting quite a few responses blocked with safety concerns, primarily because I have been using generative AI to stylize smart home announcements (i.e. ‘the coffee is ready’) with the witty, humorous, sarcastic, and occasionally insulting voice of Iron man’s JARVIS, which was formerly permissible under the Bison Model.

I’d like to do something like the following:

response = model.generate_content('[Questionable prompt here]',                                  safety_settings={'HARASSMENT':'block_none'})response.text

Perhaps this would require modification of the integration configuration?

1 Like

Did you ever find anything out with this, I am having the same problem for mostly the same reasons.

Yeah, same here. It won’t let me switch off the lights in the bedroom I guess because of it’s rather prudish nature!

ok worked it out-- should have checked first.
Settings- Devices and Services- Google Generative AI- Configure, then uncheck ‘Recommended model settings’. The dialog expands to show some new options including safety settings.

3 Likes