Mozilla has came out against Google’s push to include AI-based APIs centered around Gemini Nano. They are concerned that web developers will only design apps based on Gemini instead of choosing their preferred vendor.
Mozillans, along with other interested parties, have been mulling Google’s plans in web community discussions since the proposals started showing up last year. But Mozilla’s position shifted from "wait-and-see” to negative last week after discussions with Google developers failed to alter the Chrome team’s plans.
Google’s decision to base these APIs on Gemini Nano in Chrome means that developers will create apps based on Gemini’s behavior, Mozilla believes. Since those apps might behave differently in a different browser that used the same APIs with a different model, web developers would be likely to prefer Chrome with Gemini.
Brian Grinstead, senior principal engineer at Mozilla, has articulated his company’s position in a series of posts about the Writing Assistance APIs.
Baking Gemini into Chrome isn’t just a minor technical detail, he said in a post two weeks ago. The browser’s choice of AI model has the potential to make the user experience worse in Firefox and other browsers, as he pointed out in an example.
“We could try to work around this by shipping Chrome’s built-in model in Firefox, but that’s not how the web platform is meant to work and it points to a broader set of concerns about market competition.”
Grinstead argued web developers should be able to choose whatever AI model suits their application. Google, he contends, should not be picking favorites.