Earlier today, several people who used Gemini LLMs for AI roleplay received an email from Google. The email informed them that they have used a Google product or service in a manner that violates their Terms of Service.
Using Gemini LLMs for AI roleplay isn’t against Google’s policies, but your prompts and outputs might include content that violates their Terms of Service. And it isn’t just smut that draws Google’s attention.
Gemini API Ban Wave
The number of people who received this email grew by the hour. It was clear that Google was enforcing its Generative AI Prohibited Use Policy, and we were witnessing Gemini API’s first ban wave.
Dear Developer, We recently detected that your Google Cloud/API project Gemini API (id: …) is using a Google product or service in a manner that violates the Terms of Service for the Google API you are using and, as a result, we have restricted your use of the corresponding APIs.
The Gemini API ban wave isn’t surprising, especially if you’ve been involved in AI roleplay for some time and have seen OpenAI ban waves before. Users who used jailbreaks and generated content that broke OpenAI’s policies were banned in waves every few months.
Also Read: Why Collective Shout Is Bad News For AI Roleplay
But Google’s email shocked many, especially those new to AI roleplay and whose first experience with proprietary LLMs was with Gemini.
AI Roleplay And Google’s Generative AI Prohibited Use Policy
Using their LLMs for AI roleplay isn’t a violation of Gemini API’s Terms of Service, but generating content that involves violence or sexually explicit material violates their Generative AI Prohibited Use Policy.
Your AI roleplay doesn’t need to include just smut to attract a ban. Anything that Google views as hate speech, harassment, abuse, or violence can also put you on their radar.
Even if your roleplay messages don’t include any NSFW material, any part of your prompt, such as custom prompts, character definitions, etc., could contain content that violates Google’s policies. Furthermore, using prefills or jailbreaks to bypass Gemini’s filters violates the Gemini API Terms of Service and Google APIs Terms of Service. Doing so can result in a ban from the Gemini API.
The Services include safety features to block harmful content, such as content that violates our Prohibited Use Policy. You may not attempt to bypass these protective measures or use content that violates the API Terms or these Additional Terms.
Gemini API Additional Terms of Service
If you received an email notifying you of your Gemini API ban, you can try to appeal it. However, don’t expect a positive outcome if you used a prompt or generated content that violated Google’s policies.
What To Do Next?
Once your Gemini API ban is in effect, you won’t be able to create new projects or API keys through AI Studio.
Using multiple Google accounts to access the Gemini API also violates their policies and could cause Google to restrict your account again in the future. You’ll need to pay and use Gemini LLMs through a proxy service like OpenRouter or find another LLM to power your AI roleplay.
Also Read: DeepSeek R1 vs. V3 – Which Is Better For AI Roleplay?
If you were lucky enough to escape this Gemini API ban wave, don’t celebrate yet. You might get a similar email from Google in a few months. Only Google knows the process they use to enforce their policies, and they may have skipped over you this time.
The Gemini API ban wave may have surprised many newcomers to AI roleplay, but it also serves as an important lesson. Don’t get too attached to proprietary models. The companies behind these models can terminate your access to them at any time.
