- Alphabet was expanding its AI-generated summaries for search queries to six new countries
- Since receiving the challenge, Google has been working to enhance its AI search feature and broaden its accessibility to assist more users.
OUR TAKE
Google’s latest updates and expansions to its AI Overviews showcase both the opportunities and obstacles in incorporating AI into search. While extending the feature to more countries and improving its functionality are commendable advancements, the early inaccuracies emphasise the need for strong quality control. Adding hyperlinks and testing in-text links is a wise move, as it improves user experience while directing traffic to pertinent websites. In summary, this approach is promising but still developing, and will require ongoing refinement.
-Tacy Ding, BTW reporter
What happened
Alphabet , Google’s parent company, revealed that it is extending its AI-generated summaries for search queries to six additional countries. This announcement comes just two months after the company had to retract some of these features due to a problematic initial launch.
AI Overviews is now coming to the Brazil, India, Indonesia, Japan, Mexico and Britain, in local languages such as Portuguese and Hindi.
The search giant launched AI Overviews—featured at the top of search results pages before traditional web links—to all U.S. users in May.
Google is also enhancing the feature by adding more hyperlinks. Websites will now be displayed to the right of the AI-generated answer.
The company is also internally testing an update that would incorporate links directly within the text of the overview. This is part of an effort to “prioritise approaches that drive traffic to relevant websites,” according to a blog post released on Thursday. Budaraju said the new update would have a “three-way benefit” for Google, consumers and publishers.
Also read: Google launches Pixel 9 with new AI features
Also read: OpenAI emerges as a major threat to Google with AI search
Why it’s important
In fact, this feature faced significant criticism after screenshots of factually inaccurate answers began circulating online. These included a pizza recipe that erroneously listed glue as an ingredient and a response incorrectly claiming that former U.S. President Barack Obama is Muslim.
Google acknowledged the “odd and erroneous overviews” and announced updates to the product in a blog post in late May. These updates introduced restrictions on the types of queries that would generate AI answers and limited the use of user-generated content from sites like Reddit as source material for these answers.
“I have enough evidence to say that quality is only improving,” Hema Budaraju, a senior director of product at Google, told Reuters in an interview on Wednesday. She highlighted internal data showing that users who had access to the feature reported higher satisfaction and tended to search for longer, more specific queries compared to those who did not have access.
