TL;DR
- Google could be about to offer users new settings to control Gemini responses.
- This “content filter” would presumably offer control over the extent to which Gemini self-censors.
- The Gemini API already has settings for “damage categories” available to developers.
Artificial intelligence has made its way into virtually everything in the last year, and when it comes to AI and Google, Gemini is the name of the game. The company has been working hard to improve its models and has announced more early this month Gemini 2.0. Because of the power that AI offers, companies like Google spend a lot of time thinking about what kind of protections their AI systems should have in place. Right now, we’re investigating a new way Gemini could soon give users more control over the type of content it generates.
A Teardown of the APK helps predict features that might come to a service in the future based on current work code. However, these planned features may not be available in a public release.
Today we are reviewing beta version 15.51.24.sa.arm64 of the Google app for Android. Although this is not yet made public and live, we were able to enable a new option in the Gemini settings that appears to prepare the ability to filter Gemini output by content.
AssembleDebug / Android Authority
At the moment, selecting this option does nothing useful and only takes us to a URL for Google’s Gemini site that is not yet available. While this makes it a little difficult to say for sure how this setting is likely to work, we can draw some inferences based on what we already know about Gemini.
Google already has formal policies in place regarding what is acceptable to do with geminiand from there also has goals regarding the type of content it tries to put out. encourage Gemini to avoid. Developers, in particular, had access to reasonably accurate information. Gemini Security Settingsgiving them control over a number of “harm categories”, including hate speech, harassment and sexually explicit content.
Knowing all that is currently available, it’s possible that Google will extend some level of this same control to Gemini end users on Android and the web. That said, Google could go in a number of directions when it comes to its implementation here, from the kind of broad developer-level filters that the API currently has, to those that are perhaps even more restrictive, all in an effort to prevent Gemini from sharing anything particularly disgusting.
At the moment, it’s hard to say when this new option might go live, but we’ll keep an eye out for any further motions.
