Social media is currently flooded with examples of anomalies in Google’s new AI Overview product, ranging from advising users to apply glue on their pizza to recommending the consumption of rocks.
This problematic rollout has forced Google to manually disable AI Overviews for specific searches, leading to many disappearing shortly after being shared on social networks.
This situation is particularly surprising given that Google has been testing AI Overviews for a year. The feature, launched in beta in May 2023 under the name Search Generative Experience, has reportedly handled over a billion queries during this period, according to CEO Sundar Pichai.
Pichai has also mentioned that Google has reduced the cost of delivering AI answers by 80 percent over this time, attributing this to “hardware, engineering, and technical breakthroughs.”
However, it appears that such optimisations may have been implemented prematurely, before the technology was fully prepared.
“A company once known for being at the cutting edge and shipping high-quality stuff is now known for low-quality output that’s getting meme’d,” an anonymous AI founder told The Verge.
Despite these issues, Google maintains that its AI Overview product generally provides “high quality information” to users.”
Google has ambitious plans for AI Overviews. The current version is just a fraction of what the company announced last week, which includes features like multistep reasoning for complex queries, the capability to generate AI-organized results pages, and video search in Google Lens.
However, the immediate priority for the company is to ensure the basic functionality is reliable, a goal that is currently under scrutiny.