Google's AI is ‘hallucinating,' spreading dangerous info — including a suggestion to add glue to pizza sauce
Google's AI Overviews, designed to give quick answers to search queries, reportedly spits out 'hallucinations' of bogus information and undercuts publishers by pulling users away from traditional links.
The Big Tech giant — which landed in hot water last year after releasing a 'woke' AI tool that generated images of female Popes and black Vikings — has drawn criticism for providing false and sometimes dangerous advice in its summaries, according to The Times of London.
3 Google's latest artificial intelligence tool which is designed to give quick answers to search queries is facing criticism. Google CEO Sundar Pichai is pictured.
AFP via Getty Images
In one case, AI Overviews advised adding glue to pizza sauce to help cheese stick better, the outlet reported.
In another, it described a fake phrase — 'You can't lick a badger twice' — as a legitimate idiom.
The hallucinations, as computer scientists call them, are compounded by the AI tool diminishing the visibility of reputable sources.
Instead of directing users straight to websites, it summarizes information from search results and presents its own AI-generated answer along with a few links.
Laurence O'Toole, founder of the analytics firm Authoritas, studied the impact of the tool and found that click-through rates to publisher websites drop by 40%–60% when AI Overviews are shown.
'While these were generally for queries that people don't commonly do, it highlighted some specific areas that we needed to improve,' Liz Reid, Google's head of Search, told The Times in response to the glue-on-pizza incident.
3 Google AI Mode is an experimental mode utilizing artificial intelligence and large language models to process Google search queries.
Gado via Getty Images
The Post has sought comment from Google.
AI Overviews was introduced last summer and powered by Google's Gemini language model, a system similar to OpenAI's ChatGPT.
Despite public concerns, Google CEO Sundar Pichai has defended the tool in an interview with The Verge, stating that it helps users discover a broader range of information sources.
'Over the last year, it's clear to us that the breadth of area we are sending people to is increasing … we are definitely sending traffic to a wider range of sources and publishers,' he said.
Google appears to downplay its own hallucination rate.
When a journalist searched Google for information on how often its AI gets things wrong, the AI response claimed hallucination rates between 0.7% and 1.3%.
3 Google's AI Overviews, was introduced last summer and is powered by the Gemini language model, a system similar to ChatGPT.
AP
However, data from the AI monitoring platform Hugging Face indicated that the actual rate for the latest Gemini model is 1.8%.
Google's AI models also seem to offer pre-programmed defenses of their own behavior.
In response to whether AI 'steals' artwork, the tool said it 'doesn't steal art in the traditional sense.'
When asked if people should be scared of AI, the tool walked through some common concerns before concluding that 'fear might be overblown.'
Some experts worry that as generative AI systems become more complex, they're also becoming more prone to mistakes — and even their creators can't fully explain why.
The concerns over hallucinations go beyond Google.
OpenAI recently admitted that its newest models, known as o3 and o4-mini, hallucinate even more frequently than earlier versions.
Internal testing showed o3 made up information in 33% of cases, while o4-mini did so 48% of the time, particularly when answering questions about real people.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
31 minutes ago
- Yahoo
Week in Review: Why Anthropic cut access to Windsurf
Welcome back to Week in Review! Got lots for you today, including why Windsurf lost access to Claude, ChatGPT's new features, WWDC 2025, Elon Musk's fight with Donald Trump, and lots more. Have a great weekend! Duh: During an interview at TC Sessions: AI 2025, Anthropic's co-founder had a perfectly reasonable explanation for why the company cut access to Windsurf: 'I think it would be odd for us to be selling Claude to OpenAI,' Chief Science Officer Jared Kaplan said, referring to rumors and reports that OpenAI, its largest competitor, is acquiring the AI coding assistant. Seems like a good reason to me! Everything is the same: Chinese lab DeepSeek released an updated version of its R1 reasoning AI model last week that performs well on a number of math and coding benchmarks. Now some AI researchers are speculating that at least some of the source data it trained on came from Google's Gemini family of AI. WWDC 2025: Apple's annual developers conference starts Monday. Beyond a newly designed operating system, here's what we're expecting to see at this year's event, including a dedicated gaming app and updates to Mac, Watch, TV, and more. This is TechCrunch's Week in Review, where we recap the week's biggest news. Want this delivered as a newsletter to your inbox every Saturday? Sign up here. Business in the front: ChatGPT is getting new features for business users, including connectors for Dropbox, Box, SharePoint, OneDrive, and Google Drive. This would let ChatGPT look for information across your own services to answer questions. Oh no: Indian grocery delivery startup KiranaPro was hacked, and all of its data was wiped. According to the company, it has 55,000 customers, with 30,000 to 35,000 active buyers across 50 cities, who collectively place 2,000 orders daily. Artsy people, rejoice! Photoshop is now coming to Android, so users of Google's operating system can gussy up their images, too. The app has a similar set of editing tools as the desktop version, including layering and masking. Let's try that again: Tesla filed new trademark applications for "Tesla Robotaxi" after previous attempts to trademark the terms 'Robotaxi' and 'Cybercab" failed. Rolling in dough: Tech startup Anduril just picked up a $1 billion investment as part of a new $2.5 billion raise led by Founders Fund, which means Anduril has doubled its valuation to $30.5 billion. On the road again: When Toma's founders realized car dealerships were drowning in missed calls, they hit the road to see the problem firsthand. That summer road trip turned into a $17 million a16z-backed fundraise that helped Toma get its AI phone agents into more than 100 dealerships across the U.S. Fighting season: All gloves were off on Thursday as Elon Musk and President Trump took to their respective social networks to throw jabs at each other. Though it might be exciting to watch rich men squabble in public, the fallout between the world's richest person and a sitting U.S. president promises to have broader implications for the tech industry. Money talks: Whether you use AI as a friend, a therapist, or even a girlfriend, chatbots are trained to keep you talking. For Big Tech companies, it's never been more competitive to attract users to their chatbot platforms — and keep them there. This article originally appeared on TechCrunch at
Yahoo
44 minutes ago
- Yahoo
Sprouting Gear Inc. Founder Paul Pluss Announces Report on:
'The Unintended Consequences of the AI Race on the Livestock Industry' RAMONA, Calif., June 07, 2025 (GLOBE NEWSWIRE) -- The U.S. livestock industry, already grappling with rising feed costs and shrinking herd sizes, now faces a fast-approaching and under-recognized threat: the massive expansion of artificial intelligence (AI) infrastructure—especially data centers—and its impact on water availability, says Paul Pluss, a veteran livestock rancher and researcher focused on the intersection of agriculture, water policy, and emerging infrastructure demands. 'The water usage of data centers operated by Microsoft, Google, Meta, and Amazon remains largely unrecognized by agricultural stakeholders. Prime location for data centers is the same hot dry inland location preferred for feedlots and are often sharing the same aquifers and rivers" said Pluss. Fueled by public and private investment in AI infrastructure, the number of U.S. data centers is expected to grow from 5,426 today to more than 8,378 within five years. Many existing facilities are also expanding. These data centers—crucial for powering AI models, cloud computing, and digital services—require enormous amounts of water to cool their servers. Key figures: Each data center can consume up to 5 million gallons of water per day for cooling. Average water usage per megawatt of electricity is estimated at 6 to 7 million gallons. U.S. data center power demand is currently 35 gigawatts and rising. Annual electricity usage by data centers is expected to nearly triple, from 224 terawatt-hours today to 606 terawatt-hours within five years. Based on current and projected growth, total water use by U.S. data centers could exceed 15 trillion gallons annually—equivalent to more than 46 million acre-feet of water per year (calculated on the well-documented 5M gallons/day per center, prior to new expansions). This level of water consumption rivals agricultural water use in major farming states and could soon surpass the entire livestock industry's combined water footprint, including feed crop irrigation, drinking water, and processing needs. View the report here, as well as a articles and short videos to explain hydroponic livestock feeding and the economics behind it: The Carbon Footprint of Livestock 'Can We REALLY Slash Livestock Environmental Damage by 90 Percent?' Our Country's Water Crisis: Why Aquifers Are a Bigger Problem Than the Colorado River 'Our Country's Water Crisis' From 2 Pounds of Seed to 19 Pounds of Feed Paul PlussCEO & Founderpaul@ in to access your portfolio
New York Post
44 minutes ago
- New York Post
Why both the Left and the Right are failing American workers
Working-class voters had high hopes that Donald Trump would help them out economically: Inflation and the economy were top priorities for Trump voters in 2024. These voters have seen the fading of the American Dream first-hand: Over 90% of Americans did better than their parents in the decades after World War II, but only half born in 1980 will. Why? A major reason is that employers pocketed workers' fair share of productivity increases. Wages used to rise when productivity did; if that had continued, workers' wages would be 43% higher than they are today. 6 Only half of Americans born in 1980 are doing better than their parents, with the root causes far harder to solve than mere DOGE-style cost cuts by President Trump and Elon Musk. FRANCIS CHUNG/POOL/EPA-EFE/Shutterstock You won't hear about this from President Trump. Instead, his villain is foreign trade, and his solution is to use tariffs to bring back blue-collar jobs. But tariffs won't work quickly, because onshoring means building new factories, which takes years. And it may never happen. Meanwhile, tariffs threaten to take away the only real benefit workers got from globalization: cheap Nikes, T-shirts, and party favors. To add to that, the Trump administration is DOGE-ing away stability at the Veterans Administration and Medicaid, which middle-class and rural Americans rely on. In short, the right isn't doing right by Lunchpail Joe and Jane. But neither is the left. Instead, it's obsessing nonstop about Trump's flouting of democratic norms — that's all progressives want to talk about. Defense of democracy was a top priority for those who voted for Kamala Harris, but way, way down for those who voted for Trump. The left also wants to talk about how DOGE is firing government workers, making professionals' jobs unstable and nerve-wracking. 6 Attacks on Obama-era healthcare initiatives are another reason American workers are feeling the pinch, some critics claim.. Getty Images Welcome to our world, say non-college voters. Thirty years ago, sales personnel at Macy's had full-time jobs with benefits. Today, associates worry constantly about whether they will accrue enough hours to pay the rent, in jobs that typically lack health insurance. As a result, many workers without degrees are working all kinds of hours. A 38-year-old construction worker described the impact on family life: 'People can't get or stay married because it takes so much effort to survive. My ex-fiancée said, 'You're never around.' But I was working to get a better life for us. No one has time for their kids. It's the American Nightmare.' 6 Today, associates worry constantly about whether they will accrue enough hours to pay the rent, in jobs that typically lack health insurance. Bloomberg via Getty Images DOGE is placing college-educated professionals where workers without degrees have been for decades — worried sick about how they're going to support their families. Don't expect workers to care. In earlier eras, the left was focused on good jobs for blue-collar workers and universal programs to ensure stability for the middle class, like Social Security, Medicare, and VA home loans and college benefits. In the 1970s, the focus changed to prioritize issues of greater concern to liberal college grads: environmentalism, racism, and sexism. As political priorities changed, so did 'feeling rules' that set the parameters of our heartstrings. A good lefty should feel angst about climate change, the poor, LGBTQ+, racism, and immigrants. But blue-collar Americans who vote for Trump? They're deplorable. If you care deeply about people disadvantaged by race, gender, and country of origin — but ignore class disadvantage — then people disadvantaged by class will seek solace by flocking to those who channel their anger. In both Europe and the US, those who flock to the far right are middle-status voters in routine jobs, holding on for dear life and just waiting for the other shoe to drop. A reality check: Democrats have generally done better for working people than Republicans. Obamacare is only the most recent example. The Trump administration has cut funding to programs that underlie the stability of middle-class Americans, like the VA and Social Security, creating overly lengthy wait times. 6 'Out-classed: How the Left Lost the Working Class and How to Win Them Back' by Joan C. Williams. 6 Author Joan C. Williams. This is what the left should be focusing on, not the defense of democracy. Americans who feel they've been screwed for the last 40 years feel democratic institutions haven't delivered for them. If Democrats are seen as defending the status quo, they won't win over non-college voters who feel like the status quo isn't working for them. And nearly two-thirds of Americans lack college degrees. Without them, Democrats can't win elections. Here's the bottom line. My message for Republican powers-that-be is a question: Does the current business climate, rife with chaos, instability, and the corrosion of both democratic norms and the US credit rating, really work for you? If you'd prefer a more orderly political and business climate, you need to deliver a stable, middle-class future for Americans without college degrees. 6 The Trump administration is DOGE-ing away stability at the Veterans Administration and Medicaid, which middle-class and rural Americans rely on. Getty Images My message for Democratic powers-that-be is also a question: Do you want to win elections? If you do, you need to change policies and feeling rules to deliver both economic stability and respect for non-college voters. Because if you don't, what you now see is what you'll get. There's your coalition out of this mess. Joan C. Williams is director of the Equality Action Center at UC Law San Francisco and the author of 'OUTCLASSED: How the Left Lost the Working Class and How to Win Them Back.'
