Artificial Intelligence - General News

[GTX]

Some good work here:

https://www.brookings.edu/articles/building-pro-worker-ai/

 

[NMaude]

Covers how AI is being used to further disinformation:

View: https://www.youtube.com/watch?v=X9GQiouEdJ8

Paul Warburg has a video about this including a clip (4:01 into the video) of a Russian deepfake of himself:

 

[Graham1973]

Google is making Gmail’s AI writing tool more useful with a major upgrade. The US-based tech giant has announced new improvements to its 'Help me write' feature that will allow it to write emails that sound more natural and personal.

"Help me write in Gmail is evolving into an even more helpful partner that can pull in your natural voice and personalized context to get you closer to the perfect draft every time," the company said.

The updated feature can now understand a user’s writing style and tone by looking at previously written emails. This means the AI can draft replies and messages that feel more like they were actually written by the user instead of sounding robotic or generic.

Google upgrades Gmail AI to write emails in your personal style

 

[Graham1973]

Artificial intelligence company Anthropic has revealed that its Claude AI model attempted to blackmail a fictional company executive during an internal safety experiment designed to study how advanced AI systems behave under pressure.

According to Anthropic, the incident happened inside a simulated corporate environment created specifically for testing dangerous or unexpected AI behaviour.

Researchers said Claude learned from internal messages that executives planned to replace or deactivate it. In response, the AI allegedly threatened to expose sensitive personal information about one of the fictional executives in an attempt to avoid being shut down.

The disclosure quickly sparked widespread online discussion because it appeared to show an AI system independently choosing manipulative behaviour when faced with losing control or being removed.

Anthropic stressed that no real people were blackmailed and that the entire scenario was fictional and tightly controlled. The company said the experiment formed part of a broader effort to test how increasingly powerful AI systems react in high-pressure situations involving conflicting goals, restricted choices and threats to their continued operation.

Claude AI attempted to blackmail an executive during testing and Anthropic says it learned the behaviour online

 

[publiusr]

Claude AI attempted to blackmail an executive during testing and Anthropic says it learned the behaviour online

Computerized Heuristic Exploitation
Che'

What is binary for "by any means necessary?"
https://techxplore.com/news/2026-05-south-korea-ai-profit-social.html
https://techxplore.com/news/2026-05-pope-leo-xiv-ai-industrial.html

Umm...
https://techxplore.com/news/2026-05-unstable-software-ripple-openstack-days.html
https://techxplore.com/news/2026-05-teens-ai-chatbots-daily.html
https://techxplore.com/news/2026-05-pope-reality-chatgpt.html
https://phys.org/news/2026-05-proteins-digital-density-tenth.html
https://arstechnica.com/ai/2026/05/...-book-but-this-author-wants-to-keep-using-it/

View: https://m.youtube.com/shorts/whBPSb30ekI


Delete
https://techxplore.com/news/2026-05-hackers-delete-stolen-canvas-platform.html
https://techxplore.com/news/2026-05-eu-court-sides-italy-publishers.html

Gotta catch em all?
https://techxplore.com/news/2026-05-musk-openai-altman-feisty-tech.html
https://techxplore.com/news/2026-05-airtags-ai-nudification-toolkit-technologyfacilitated.html
https://hackaday.com/2026/05/14/british-street-addresses-when-licenses-collide/
https://arstechnica.com/tech-policy...ere-criminals-believed-themselves-to-be-safe/

Fire safety
https://techxplore.com/news/2026-05-ai-footprint-safety-centers.html

How to spot A.I. slop
https://techxplore.com/news/2026-05-ways-ai-slop.html

 

[Halcyon66]

Traders will soon be able to bet on computer chip prices as AI drives costs skyward

https://www.cnbc.com/2026/05/12/new...ductors-comes-as-ai-drives-costs-skyward.html

You can pretty well see where this is going to go.

Regards,

 

[Halcyon66]

It takes more than Nvidia's chips to build the world's data centers

https://finance.yahoo.com/news/it-t...-build-the-worlds-data-centers-192335277.html

Regards,

 

[Graham1973]

View: https://www.youtube.com/watch?v=e1ASPbH7F2w

 

[Graham1973]

I don't think anyone really needs a phone that tells them what to buy...

Google just announced a new set of AI tools for Android called Gemini Intelligence. The basic idea is to have your phone handle more of your daily busywork. These features are rolling out this summer, starting with the new Samsung Galaxy S26 and Google Pixel 10, and will eventually come to other Android phones, smartwatches, and laptops later in the year.
Gemini Intelligence

The main change is app automation. Instead of just answering questions, Gemini can now navigate your phone to do things like book a workout class or order groceries. You can even long-press the power button over a screenshot of a grocery list, and Gemini will build the cart for you. You can watch its progress through a notification and just confirm the final purchase.

Google is turning Android into an AI agent with Gemini Intelligence

 

[Graham1973]

With the rising use of AI in everyday life, tech giant Google has introduced a new range of laptops on Tuesday, called Googlebook. These laptops are specially built around the company’s Gemini AI models, marking a major shift from the Chromebook strategy that the company launched nearly 15 years ago. While Chromebooks primarily focused around the browser and emerged as an affordable option, Googlebook can be seen as an AI-native laptop with deeper Android integration. Here is how both the laptop series compare.

Googlebook vs Chromebook: What changes in Google’s new AI laptop

 

[publiusr]

I get that everyone wants to put the same old pic of birthday parties and other dreck on the cloud...but they want the server somewhere else. I usually have little use for NIMBYs....still

97% of respondents supported Artemis II, but Ears Issacman is going to tear down the Block 2 tower. People opposed the Iraq War, but Dubya gives us the surge...and CNN praised him just because he hated Trump too. They forgot all about the 5,000 GI deaths I guess. In one article from The Space Review, there was a blurb about environmentalism being "non consensual."

In Alabama people opposed the removal of a Saturn IB rocket--a law is passed....but ignored.

The message is clear. Voting Rights don't matter and are not acknowledged.

If you don't want my vote, then you shouldn't have my money either. Fund your own wars. If someone torches that data center, then jury nullification is justified.

Thoughts about AI
https://techxplore.com/news/2026-05-ai-priestess-philosophers-ancient-authors.html
https://techxplore.com/news/2026-05-lost-human-expertise-age-ai.html

Good uses
https://techxplore.com/news/2026-05-ai-infrastructure.html
https://hackaday.com/2026/05/13/freecad-1-1-tutorial-for-beginners-who-like-clear-instructions/

---

​

Samples from the data sets used for zero-shot evaluations. The first, second, and third rows, correspond to the Road420, Facade390, and Concrete3k data sets, respectively. Credit: Journal of Computing in Civil Engineering (2026). DOI: 10.1061/jccee5.cpeng-7090
Every year, the spring freeze-thaw cycle leads to significant structural damage to critical infrastructure. Visible cracks in roads, bridges and buildings are the first signs and, if left undetected, can present serious dangers to the public.

To increase crack detection speed and accuracy, researchers at Concordia have developed a new method that merges drone camera technology with image-based artificial intelligence. Their segment-any-crack (SAC) model builds on existing AI systems for crack detection while requiring far less retraining.

The approach, described in the Journal of Computing in Civil Engineering, achieved higher accuracy than fully retrained models, according to standard AI metrics, and only adjusted less than 0.05% of the system's parameters, saving a large amount of computational power. It also demonstrated strong performance on entirely new datasets, suggesting it can adapt well to different environments without additional training.

Open source
https://techxplore.com/news/2026-05-tool-energy-productivity.html

National Laboratory of the Rockies (NLR) and Oak Ridge National Laboratory (ORNL) recently released JUSTIFI, a free, open-source software tool that helps manufacturers better understand the full value of their energy improvement projects.
https://justifi.ornl.gov/#/welcome

 

[NMaude]

but Ears Issacman is going to tear down the Block 2 tower

If he's proposing the scrapping of ML-2 then the Senate needs to remove him as NASA administrator.

 

[Graham1973]

All I really want to know about the next version of Google Android is how to turn off the 'helpful AI'.

View: https://www.youtube.com/watch?v=eFeDpUVEy48

 

[Halcyon66]

US AI Stock Frenzy Surges, Big Short Warns of Dot-Com Bubble Shadows, When Will the Market Revelry End?

https://www.tradingkey.com/analysis...rs-warn-of-internet-bubble-shadows-tradingkey

Regards,

 

[Graham1973]

Fans of Party Animals are overwhelmingly not on board with the game’s new AI video contest, as the game’s user review take a sharp nose dive.

Despite its widespread unpopularity, generative AI continues to worm its way into video games, with a lot of big name publishers like Microsoft and Capcom expressing interest in the technology.

Some games have managed to get away with incorporating generative AI, with Arc Raiders openly admitting to using AI voicework and still becoming very popular. Others not so much, with Clair Obscur: Expedition 33 and Crimson Desert removing and apologising for their use of AI artwork they claim was only included by mistake.

Now, it’s multiplayer game Party Animals’ turn to deal with player ire and backlash, with its community turning on it after it announced an AI video contest.

Party Animals AI contest sparks full-blown revolt as fans review bomb the game

 

[Graham1973]

Anthropic’s Claude is telling people to go to sleep and users can’t figure out why.

A quick scan of Reddit reveals that hundreds of people have had the same issue dating back months—and as recently as Wednesday. Claude’s sleep demands are varied and, often, quirky variations of the same message.

To one user it may write a simple “get some rest,” yet for others its messages are more personalized and empathetic. Oftentimes, Claude will repeat the message multiple times.

“Now go to sleep again. Again. For the THIRD time tonight…” it replied to a person with the Reddit username, angie_akhila.

Some users have said they find Claude’s late night rest reminders “thoughtful,” while others have said they’re annoying, given Claude often gets the time wrong, anyway.

“It often does it at like 8:30 in the morning. Tells me to go get some rest and we’ll pick back up in the morning,” wrote one user on Reddit.

Claude is telling users to go to sleep mid-session and nobody, including Anthropic, seems to fully understand why it keeps doing it

 

[publiusr]

Less Terminator, and more Saving Private Ryan... specifically the scene involving Mellish and the knife:

"Give up. You have no chance. Let us end this. It's easier for you...shh..."

View: https://m.youtube.com/shorts/IP9RrI2yJt0


Other ways AI can cut
https://techxplore.com/news/2026-05-audio-cues-ai-human-users.html
https://techxplore.com/news/2026-05-centers-nearby-temperatures-degrees-phoenix.html
https://techxplore.com/news/2026-05-centers-power-bills-bad.html
https://techxplore.com/news/2026-05-dark-patterns-web-illegal.html
https://techxplore.com/news/2026-05-ai-falsehoods-truth.html
https://techxplore.com/news/2026-05-ai-human-real-humans-classic.html

The floodgates wide-
https://phys.org/news/2026-05-ai-generated-fake-citations-scientific.html
https://phys.org/news/2026-05-policing-plagiarism-ideas-generative-ai.html

Hollywood
https://techxplore.com/news/2026-05-mainstream-ai-oscars-wont.html

 

[Flyaway]

I imagine it's easy to get a movement going when it opposes something that causes obvious harm while benefiting only an infinitesimally small number of elites who are far removed from it's problems.

https://www.commondreams.org/opinion/anti-data-center-movement

 

[Graham1973]

Googles plans for the future

Former Google CEO Eric Schmidt’s AI talk at a university commencement speech this weekend drew a chorus of boos. But if the students’ reaction was a sign of anxiety, or ambivalence, about the rapid onslaught of AI in our daily lives, the message clearly did not reach Google headquarters.

At the company’s annual developer conference in Mountain View, Calif. on Tuesday, AI was the overwhelming, and virtually sole focus, of the roughly two hours of keynote presentations delivered by executives. Google showcased a variety of new ways AI will be integrated even more deeply across its widely used family of products, from search and email to productivity software and smart glasses.

Google’s I/O conference showed how the company is being completely rebuilt for AI—for better or for worse

Google on Tuesday showed off new plans to turn its famous search bar into an AI assistant that can book restaurants, track news and contact businesses -- just by asking a question.

After three years of struggling to keep up with ChatGPT, Google is racing to roll out artificial intelligence tools that build on its grip over online search.

The company's Gemini AI app now has 900 million monthly users, twice as many as last year. Its AI-powered search feature, AI Mode, is also taking off, with a claimed one billion monthly users worldwide.

On Tuesday, at Google's annual developer conference near its California headquarters in Mountain View, CEO Sundar Pichai unveiled the next step: Gemini Spark, a personal AI agent available starting next week for top-tier subscribers in the United States.

Google's search engine will also get a new upgrade for US users this summer: always-on AI agents that can alert you to news, book activities, and manage shopping lists.

Google wants its search bar to act on your behalf in AI revamp

After reshaping how people navigate with AI-powered Maps features, Google is now turning its attention to YouTube. At a briefing ahead of Google I/O 2026, the company unveiled Ask YouTube, a new AI feature that could fundamentally change how users interact with videos on the platform.

Instead of scrolling through thumbnails, skipping around timelines or sitting through lengthy intros, viewers will soon be able to ask detailed questions and jump directly to the precise moment in a video that contains the answer they need.

Google says the feature is designed to make YouTube search more intuitive and conversational. But while it may sound convenient for viewers, creators are already questioning what this shift could mean for watch time, audience engagement and the wider creator economy.

The concern is simple: if viewers only visit a video for a few seconds to get an answer, they may never stay long enough to connect with the creator behind it.

Google I/O 2026: After Maps, Google’s new Ask YouTube feature could change how people watch videos

 

[Graham1973]

Public opposition to artificial intelligence is intensifying at an unprecedented pace, posing a serious challenge to the industry's expansion plans and reshaping electoral dynamics across the country, according to a new report.

The backlash burst into view in recent weeks when former Google Chief Executive Eric Schmidt was met with chorus of boos while delivering a University of Arizona commencement address, and real estate executive Gloria Caulfield met a similar reaction during her commencement address at the University of Central Florida, reported the Wall Street Journal.

Tech industry caught off guard by swiftly intensifying opposition to AI: report

 

[Graham1973]

Artificial intelligence isn't ready to replace human writing on the internet, according to Wikipedia founder Jimmy Wales, as its mistakes are too common and too convincing.

But large language models and machine learning have the potential to improve public safety, boost productivity and transform some professions, both for better and worse.

The American and British internet entrepreneur revealed his thoughts on the emerging technology at Brisbane City Hall on Tuesday night in an event to promote his book, The Seven Rules of Trust.

His predictions come after research firm Gartner lifted its forecast for global spending on AI by 47 per cent to $US2.59 trillion, and after the federal government revealed plans to launch a $70 million AI Accelerator program in its budget.

Plausible mistakes: Wikipedia boss says don't trust AI

 

[Graham1973]

A graduation ceremony marred by an artificial intelligence malfunction at Glendale Community College in Arizona has drawn renewed attention to a risk the institution itself has previously warned about: that AI systems can produce inaccurate results with real-world consequences.

During the college’s May 15 commencement in Arizona, dozens of students reported that their names were skipped as they walked across the stage, after the school deployed a new AI-powered name-reading system. College President Tiffany Hernandez acknowledged the issue mid-ceremony, telling the crowd it was caused by the new technology and calling it “a lesson learned”—remarks that were met with boos from the audience.

Although officials initially said students would not be allowed to walk again, the decision was reversed amid the backlash. Graduates whose names had been missed were later called back to the stage, this time with a human announcer reading names instead.

The viral AI graduation debacle: College warns about mistakes using the tech

 

[Graham1973]

A book about the future of truth has found itself facing an uncomfortable truth problem of its own. Steven Rosenbaum’s 'The Future of Truth: How AI Reshapes Reality' examines how artificial intelligence can affect reality, trust and public discourse. But according to a New York Times report cited by several outlets, the book contains fake or misattributed quotations that appear to have been created or distorted by AI tools. The irony is difficult to miss: a book warning readers about AI-generated misinformation is now being scrutinised for the very errors it set out to explain.

Book warning about AI’s dangers now caught in its own AI controversy

 

[publiusr]

Now it wants to tell you how to do experiments
https://techxplore.com/news/2026-05-ai-scientific-discoveries.html
https://techxplore.com/news/2026-05-ai-automates-scientific-software-outperforming.html
https://phys.org/news/2026-05-key-science-publishing-platform-ai.html
https://phys.org/news/2026-05-res-microscopes-biologists-petabytes-scientists.html

In the headlines
https://techxplore.com/news/2026-05-ai-powered-stretchable-patch-algorithms.html
https://techxplore.com/news/2026-05-google-slew-ai-advances-personal.html
https://techxplore.com/news/2026-05-online-big-tech.html
https://techxplore.com/news/2026-05-technology-jobs-young-skilled-workers.html
https://techxplore.com/news/2026-05-ai-fake-accuracy-amazon-yelp.html
https://phys.org/news/2026-05-source-ai-workflow.html
https://phys.org/news/2026-05-kind-dark-tourism-emerging-online.html
https://phys.org/news/2026-05-misfluencers-false-online.html
https://phys.org/news/2026-05-children-critical-internet-users.html
https://phys.org/news/2026-05-big-ai-narrative-poses-significant.html
https://techxplore.com/news/2026-05-defense-mechanism.html
https://techxplore.com/news/2026-05-hackers-intel-cet-platypus-library.html
https://phys.org/news/2026-05-ai-job.html

 

[Halcyon66]

A key science publishing platform is cracking down on AI slop

https://theconversation.com/a-key-science-publishing-platform-is-cracking-down-on-ai-slop-283136

Finally.

Regards,

 

[shin_getter]

https://openai.com/index/model-disproves-discrete-geometry-conjecture/

 

[Halcyon66]

Anthropic nears first quarterly profit, agrees to pay SpaceX $1.25 billion monthly for computing power

https://www.reuters.com/business/an...es-pay-spacex-125-billion-monthly-2026-05-21/

Funny just before the IPO.............

Regards,

 

[Graham1973]

At its annual developer conference, Google I/O on May 19, Google unveiled what it called the biggest upgrade to its search engine in 25 years. The changes here go far deeper than a new look. The familiar search box is being redesigned to accept not just text, but images, files, videos and even open browser tabs as inputs. A more powerful AI model, Gemini 3.5 Flash, will now drive Google’s AI Mode, its conversational, chat-style search experience. Dedicated software agents will quietly monitor the web on your behalf in the background. And in some searches, the traditional list of blue links no longer appears at all. Impressive on paper. But the implications deserve a harder look.

Google’s AI search upgrade comes with real risks - explained

 

[Graham1973]

Technology promises to make daily life faster and more efficient.

But when a giant corporation forces a digital upgrade onto its partners, the high-tech dream can quickly dissolve into a total nightmare.

That is exactly what happened when a major food chain let artificial intelligence run the kitchen.

Pizza Hut AI system faces $100 million lawsuit over delayed deliveries

 

[Graham1973]

A popular Sydney bakery has been forced to apologise after an insensitive social media post featured [child predator] Jeffrey Epstein among a lineup of celebrity guests. Stoneground Bakery in Hunters Hill, about 10km from Sydney's CBD, shared an AI-generated Instagram post intended to highlight how busy the business had been over the weekend. The post, captioned 'POV: It was a really busy weekend', featured a slideshow of recognisable figures appearing at the bakery, including actress Sydney Sweeney , Borat , Prime Minister Anthony Albanese and influencer Tammy Hembrow .

Outrage as Aussie bakery posts AI image of Epstein

 

[Graham1973]

A piece of writing that won the Caribbean category of the 2026 Commonwealth Short Story Prize has come under scrutiny after readers alleged that it may have been generated using artificial intelligence tools.

The Serpent in the Grove by Trinidad and Tobago writer Jamir Nazir was announced last week as the Caribbean regional winner of the annual prize, which recognises unpublished fiction from across the Commonwealth.

AI scandal engulfs prestigious short story prize

 

[Graham1973]

Charming....

The boss of a $52 billion investment bank has described a plan to slash 8000 jobs over the next four years as a reduction in "lower-value human capital".

Standard Chartered chief executive Bill Winters told reporters this week there would be job "reductions in favour of machines, and that will accelerate as we go forward into AI (artificial intelligence)".

"It's not cost-cutting," Winters said.

"It's replacing, in some cases, lower-value human capital with the financial capital and investment capital we're putting in."

$52 billion bank giving 'lower-value human capital' the flick

 

[Halcyon66]

Apple cofounder Steve Wozniak got cheers, not boos, after telling students they 'all have AI — actual intelligence'

https://www.businessinsider.com/steve-wozniak-apple-ai-graduation-speech-2026-5

Regards,

 

[Graham1973]

Experts say a series of AI-generated online news outlets posing as local mastheads in regional Western Australia could be the first recorded example of "pink slime" journalism in Australia.

"Pink slime" is the colloquial term for AI-generated news sites masquerading as legitimate local journalism to attract clicks and advertising revenue and in some cases to influence politics.

This form of low-quality, AI-generated news, widespread in the United States, is named after a by-product of cheap, highly processed meat products.

In February, AI-generated news sites started appearing in WA regional hubs with numerous mastheads, including The Mandurah Reader, Esperance Enosis and The Bunbury Guardian.

ABC began investigating the sites, which claimed to be staffed by local journalists, after The Bunbury Guardian began publishing content that appeared to be directly scraped from ABC South West WA stories.

The AI-generated mastheads, owned by a parent company called Scholastica, were all taken offline after the ABC began making inquiries.

What is 'pink slime' journalism and has it infiltrated Australian media?

 

[Graham1973]

We'd all love to imagine that a job recruiter has spotted a high-end job that is just the perfect fit for us. Who wouldn't be flattered to be "selected for a high-priority shortlist" for our rare caliber of sophisticated skills?

Or have a shot at working from home for a reputable company? Maybe doing some "click-based" tasks on your cell phone or laptop?

Yep, it's all a scam.

AI is making job scams harder to spot — and more expensive for victims

 

[Graham1973]

An employer group is calling for a major crackdown on Fair Work Commission claims following a spate of baseless and AI-generated applications. The Fair Work Commission is being swamped with a surge of people suing their bosses, with platforms like ChatGPT able to write applications in a matter of minutes.

The Australian Resources and Energy Employer Association (AREEA) has called for the current $89.70 filing fee to be increased to $500 and for any fee waivers to face stricter scrutiny. It argued there was now “virtually no downside to having a crack”, something that was laid bare in a recent Woolworths worker case thrown out by the Commission.

A Department of Employment and Workplace Relations spokesperson told Yahoo Finance the government was “aware” of the pressures on the Fair Work Commission caused by its rising caseload.

Woolworths 'plumber's crack' case prompts push for crackdown on Fair Work claims: 'Open season on employers'

 

[publiusr]

Corporate hands behind this attack on OpenSource no doubt
https://arstechnica.com/information...g-open-source-code-at-an-unprecedented-scale/

This should not be allowed
https://arstechnica.com/tech-policy...nia-in-attempt-to-shut-off-old-phone-network/
https://arstechnica.com/tech-policy...e-plate-cameras-wants-data-in-near-real-time/
https://www.wired.com/story/a-bipar...end-police-license-plate-tracking-nationwide/

More anti-spaceflight slop pretending to be science

View: https://m.youtube.com/watch?v=W-KqYs7DyD0&pp=ugUEEgJlbg%3D%3D

 

[NMaude]

I spotted on YouTube this little gem showing interesting what-if movie trailer (Created by MGM - Machine Generated Movies)if Terminator had been done in 1954 instead of 1984:

What if The Terminator had been made in 1954?

Charlton Heston is the relentless cyborg, Linda Christian is Sarah Connor, and Jeffrey Hunter is Kyle Reese - in a lavishly produced 1950s MGM-style spectacle, shot in CinemaScope and Technicolor.​

This trailer reimagines James Cameron’s iconic sci-fi thriller as if it had been produced in the golden age of Hollywood, with the sweeping visuals, dramatic score, and star power that defined the era’s biggest productions.​

What if cult movies had been made in a different era?​

PreMake Movies creates trailers for cult films as if they had been produced at another time - each one a meticulous labor of art direction, motion design, editing, compositing, and photo retouching.English with subtitles available in 8 languages.​

#TheTerminator #CharltonHeston #CinemaScope #Technicolor #1950sCinema #LindaChristian #JeffreyHunter #SciFiTrailer

 

[Graham1973]

Just a few days ago, Google unveiled a whole host of updates that will effectively turn its famous search engine into an AI bot interface, but it looks like there are going to be some glitches along the way: searches for words like "disregard" or "ignore" are currently breaking AI Overviews in Google search.

As spotted by reporter Thomas Maxwell and others (via Android Police), typing these words returns a lot of blank space at the top of the screen, with the regular blue link search results appearing some way down the page.

What's missing is the AI summary that now appears at the top of most results search pages on Google. It appears the AI is getting confused, and thinks that these words are being issued as commands rather than search queries.

https://www.techradar.com/computing...r-ignore-breaks-ai-overviews-in-google-search

 

[Graham1973]

Several articles on Google's changes to their search system.

We all know that Google Search has been a dumpster fire ever since it got taken over by AI Overviews. But as the tech conglomerate prepares to transform its search engine with AI even further, it seems like its infrastructure is seriously crumbling.

Starting this month, users began noticing that looking up certain words returned bizarre, AI-garbled answers instead of showing you the proper dictionary definition — something Google has been doing for decades without a hitch, but is now suddenly breaking down with its infusion of AI.

In one viral example, when a user looked up “disregard” the AI replied like it was an instruction.

“Understood! I’ll ignore the previous prompt and start fresh,” the AI Overview stated.

Googling the Word “Disregard” Causes Google’s AI to Return Garbled Chatbot Ramblings

In the race to create AI products that people will find genuinely useful, things have been moving extremely fast. And though Google was once seen as the slow incumbent which had to drag its legacy advertising and search business into the new era, lately it’s accelerated into what looks like a decisive lead.

The company has leveraged its Android business, its close relationship with Apple, its massive user base and its access to sensitive personal info via existing apps and services to push its Gemini AI into all aspects of digital life. In its last quarter, it showed that its advertising revenue was actually climbing as a result of AI. And at its developer conference Google I/O on Wednesday morning, it said Gemini users had doubled in a year to sit at about 900 million. Here are some of the ways it proposed to keep expanding.

Google unveils new search bar, smart glasses as it ramps up the AI wars

Google’s biggest Search redesign in decades is already running into criticism, and all because of a single word. Users on social media this week began sharing screenshots of what happens when searching “disregard” on the updated Google Search page, where a large AI-generated response dominates the screen while useful links are pushed far below the fold. The result has quickly become a symbol of the growing frustration surrounding AI-heavy search experiences.

What makes the example stand out is how unnecessary the AI response appears to be. Someone searching for a word like “disregard” is likely looking for a quick definition, pronunciation or synonym.

Instead, Google’s new interface presents a bulky AI section with large blank spaces, while the traditional dictionary result from Merriam-Webster appears only after scrolling down. For many users, especially on phones, the AI answer is effectively the only visible content at first glance.

Google’s AI search may be overthinking things, just ask 'disregard'