Google is planning to put limits on its AI overviews, an artificial intelligence search experience, following uproar over inaccurate results to consumers. After almost a year of experimentation, the tech giant, at its annual developer conference ‘Google I/O 2024’ on May 14, had made the announcement regarding the launch of AI Overview. This feature was said to be aculmination of Google’s Search Generative Experience (SGE).
In what can be called bizarre, the AI Overview’s responses went viral earlier this week as it asked users to use glue to make cheese sticks on a pizza or eat rocks. The screenshots of misleading results have been widely shared across the social media.
In a blog post, Liz Reid, VP, head of Google Search, called out data voids for the inaccurate results. Reid clarified that while the AI Overviews don’t generally hallucinate, they may end up misinterpreting what is already on the web.
“This included robust red-teaming efforts, evaluations with samples of typical user queries, and tests on a proportion of search traffic to see how it performed. But there’s nothing quite like having millions of people using the feature with many novel searches. We’ve also seen nonsensical new searches, seemingly aimed at producing erroneous results,” Reid added in the post.
Following the launch, “hundreds of millions” of users in the United States accessed AI Overviews, which was to be rolled out to a billion people by this year’s end. It enabled users to get an AI-generated summary of any topic, providing links to further delve deep into the queries being searched for.
The launch, however, came amid intensifying competition in the generative AI era, with OpenAI, Microsoft upping the ante, with Google seeking to reimagine its flagship search product.
Nonetheless, Google has admitted that it will make improvements, accepting that some questions could not be answered due to their satirical nature.
+ There are no comments
Add yours