It’s a move in the right direction for AI transparency.
Google has called Bard an “experimental conversational AI service” since first announcing it earlier this year. The company says it’s constantly tweaking its models, but without a central place to learn what’s changed, it can be hard to know what it’s actually capable of. In an attempt to be more open about Bard’s development, Google has created a new “experiment updates” page where anyone can find information on recent updates to Bard, including new features and bug fixes.
So far each update posted includes a “what” and a “why.” For the updates page itself, Google says it was created so “people will have an easy place to see the latest Bard updates for them to test and provide feedback.”
Google added two other announcements in the same post. The first one explains there are now additional search options when a person clicks “Google it.” The other is an update to Bard’s math and logic skills, with Google explaining that Bard isn’t always giving the right answers there — something the company is attempting to fix. This improvement comes shortly after Google switched Bard to a more advanced language model.
While the breakdowns are pretty short and vague as to exactly how Google is creating these updates, it’s definitely a start in the right direction for giving the public more insight into their thought process. Plus, at a time when both industry professionals and people in general are openly worrying about unregulated advancements in AI technology, Google’s transparency is likely to win them a few trust points.