Google is expanding its AI Overviews feature, which summarizes complex questions at the top of traditional search results, to six additional countries: India, Japan, Mexico, Indonesia, Brazil, and the United Kingdom.
This expansion, effective from Thursday, August 15, includes support for local languages alongside English. The move comes less than three months after the feature’s initial launch in the United States, where it garnered attention for unusual suggestions, such as advising users to eat rocks or apply glue to pizzas.
Challenges in Expanding AI Overviews to Global Markets
The global rollout raises concerns about the potential for similar issues in other languages. In an Engadget interview, Hema Budaraju, Senior Director of Product Management for Search at Google, acknowledged the challenges involved, particularly the complexity of ensuring quality across the web in multiple languages while integrating large language models (LLMs).
To mitigate such risks, Google has conducted language-specific testing and employed red-teaming techniques to stress-test the system against potential misuse.
How Is Google Addressing Past AI Mishaps?
In response to the earlier incidents, Google has implemented additional safeguards, including limiting the inclusion of satire and humor in AI Overviews and restricting the types of queries that can trigger the feature. Moreover, to enhance transparency, Google will now prominently display source links on the right-hand side of AI-generated answers, with some users also seeing links embedded directly within the text of the summaries. This move aims to address concerns from publishers about reduced traffic to their websites due to AI-generated content.
Additionally, users who opt into Google’s Search Labs platform can experiment with new features, such as saving specific AI Overviews for future reference and simplifying the language of AI-generated answers.
