At the Google I/O event, the company announced a new search feature called “AI Insights.” Since its launch, this feature has become the talk of the town due to its strange and misleading responses. As people began to have doubts about Google’s AI accuracy and efforts, the company finally decided to address the growing concern and address users about how it works and why. for which it generated such dubious responses. Find out what Google said about AI preview incidents.
Google defends AI previews
On the blog job, Google first highlights how AI previews are gaining traction and how people appreciate its use and capabilities. Google said: “People are happier with their search results and are asking longer, more complex questions that they know Google can now help them with. » Later, the blog highlighted the strange insights and why it is so important to address and clarify why they generate such misleading results. Before we get into the topic, Google briefly explained how AI Previews work.
Read also : Google AI Overview Generates Misleading Answers, Tells Users to Eat Rocks, Glue
Google has clearly mentioned that AI insights do not generate results based on training data, but are powered by a custom language model integrated with major web ranking systems. This means that the feature selects information from relevant or high-quality results. Google said: “AI previews generally don’t hallucinate or make things up like other LLM products would. »
On the other hand, he also points out that AI previews generated strange responses due to “misinterpretation of queries, misinterpretation of a nuance of language on the web, or lack of ‘information available’.
Read also : Introducing Google Search AI: how to bypass them and see the results directly
AI provides insight into strange responses
In the blog, Google responds to answers that have gone viral on social networks, such as “How many stones should I eat” or “Use glue to make cheese stick to pizza.” The company said the feature is not capable of generating responses for “satirical content” due to a lack of information. Therefore, the features link to websites that offer such query answers or provide reliable first-hand knowledge.
Read also : Google Password Manager Now Gets Family Password Sharing Feature: What It Is and How It Works
However, Google also acknowledged that AI previews misinterpret the language of web pages and have incorporated improvements for these queries.
Improvements to AI previews
The company mentioned having made a dozen technical improvements to avoid such incidents:
- Google added an improved detection mechanism for absurd queries and restricted satirical and humorous content.
- Google has limited the use of user-generated content (UGC) in answers that could provide misleading advice.
- It also added restrictions on queries where AI insights are not helpful or could generate accurate or informative answers.
- For news and health-related topics, Google will completely limit the display of AI previews due to the dynamic nature and events.
One more thing ! We are now on WhatsApp channels! Follow us there to never miss any updates from the tech world. To follow the HT Tech channel on WhatsApp, click on here join now!
