Because generative AI is sometimes great and sometimes wrong.
Artificial intelligence-powered systems have been growing at high speed in recent years and while they are incredibly powerful, there’s still something about them that leaves us wanting more. Perhaps more creativity? Human connection? Oh, of course, research.
Just take a look at Google Search, which has been using its Gemini AI model to provide AI-generated answers and summaries to search queries. Users have noticed, though, that these summaries or “AI Overviews” are sometimes great and sometimes wrong, leading you astray with poor information from bad sources.
If you, like plenty of people, are interested in disabling AI in Google Search, there is some good news. Turns out, it is damn easy. One simply has to drop an F-bomb or other expletives in their search query, and Google will not return with AI Overviews at the top of the results page.
Related story: Meta is done with fact-checking: What this means for Filipino users
It is unclear who initially discovered this quirky yet useful technique, but tech publication Ars Technica recently reported that it has been going around social media for quite some time now. True enough, our test confirmed that adding some casual swear words to your search queries can turn off Google AI Overviews like a switch.
Let’s say you search for “creative ways to save money for travel,” the results page will present a blurb or AI Overviews above the blue links related to your query. However, if you type “F*cking creative ways to save money for travel” on Google’s search bar, you will instead be given a standard list of blue links without the AI-generated summary.
The same goes when you search for “How to turn off Google AI Overviews” on the search engine. The results page will show AI Overviews, telling you that “there is no direct way to turn off” the feature, but you can “hide” it by “selecting the ‘Web’ filter after performing a search.” But if you instead ask Google “How to f*cking turn off Google AI Overviews,” you will be given a set of useful web links without any AI-generated summary at the top.
When Google launched its AI Overviews in December 2023, it went viral for its nonsensical—even laughable—answers to queries. One bizarre query that got everyone talking was on how to keep your cheese from sliding off your pizza. While most of the answers on Google’s AI Overviews seemed normal, what got everyone shaking their heads was the system’s suggestion to “mix about 1/8 cup of non-toxic glue to the sauce to make it tackier.”
There was also one instance when AI Overviews recommended “eating pebbles” to improve gut health. According to a report by tech outlet Gizmodo, it is believed that Google sourced the information from comments on the popular online community platform Reddit. The result also shows that AI doesn’t know how to differentiate sarcasm or satire from facts.
Related story: Is Blue Sky the new Twitter? Why everyone’s leaving X for a ‘brighter’ social media scene
Google is not the only tech firm facing backlash for what people call “AI hallucination.” Microsoft’s Bing chatbot was also proven to “struggle with accuracy” based on a test conducted by CNN in 2023. At that time, the outlet asked Bing for “Meta’s fourth quarter results.”
Although Bing gave a response that said, “according to a press release,” the information it listed as bullet points was incorrect. The search engine stated that Meta generated $34.12 billion in revenue when the actual amount was only $32.17 billion. It also said that the company’s revenue grew from the previous year when, in fact, it had declined.
Meanwhile, including swear words in your search seems to be a goofproof way to knock off Google Search’s AI Overviews feature because the tech giant has likely trained Gemini to avoid using profanities. And, as a quick workaround, the feature is altogether disabled when expletives are present in queries.
It’s worth keeping in mind, though, that adding curse words to your search can significantly change the web link results. Don’t be shocked if the results turn out as spicier as expected, especially if SafeSearch is turned off.
At the very least, you can vary the choice of curse words you’ll use for your searches. If you don’t want a curse-filled search history, savvy searchers suggest adding “&udm=14” to the search URL to help you get rid of both the AI Overviews and Web Snippets. You can also type “-ai” (without the quotes) at the end of your Google searches to turn off the feature.
Related story: 14% of Filipino workforce at risk of losing their jobs to AI. Here’s how to protect your career
