Google admits that AI Overviews have many limitations, and users play an important role in improving them

Google admits its new AI Overviews feature has problems. After a series of false and humorous information appeared on social networks, Google officially apologized.

Instead of providing accurate information, AI Overviews sometimes give meaningless or misleading answers. Google explains that the system does not “invent” information, but makes mistakes due to “misinterpreting search context, misunderstanding internet nuances, or not finding enough quality sources.”

Google also said some screenshots circulating online were fake, while some search questions were meaningless, such as “How much ice is enough to eat?”. Due to the lack of factual information on this topic, AI Overviews directed users to humorous content (a site of a geology software provider).

What’s worrying is that Google claims to have thoroughly tested this feature before launching it. Many people question whether Google ignores queries that may produce misleading results.

Besides, Google downplays the importance of Reddit data in the knowledge source of AI Overviews. Although users often add “Reddit” to their searches, Reddit is not a reliable source of information. AI Overviews can point to Reddit posts to answer questions without distinguishing between actual knowledge and what is a joke.

Google admits “online forums are often a good source of information, but can also sometimes lead to unhelpful advice, such as using glue to get cheese to stick to pizza.”

Google AI overview suggests adding glue to get cheese to stick to pizza, and it turns out the source is an 11 year old Reddit comment from user F*cksmith 😂 pic.twitter.com/uDPAbsAKeO

— Peter Yang (@petergyang) May 23, 2024

Although Google has suffered setbacks in the past week, they are quickly improving the situation. The company said it has identified AI Overviews error patterns and will focus on building a mechanism to detect meaningless queries, limiting the use of user-generated content to possible answers. misleading, limits AI Overviews to non-breaking news topics that require high accuracy, and adds protections for health searches.

With the rise of AI chatbots today, the question is not whether they will surpass Google Search in helping us understand the world’s information, but whether Google Search can keep up with AI trends to compete with them or not. Although Google’s mistakes may seem silly, it’s still too early to knock Google out of the race, especially since Google has a huge team of beta testers – all the users searching for products. day.