惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

宝玉的分享
宝玉的分享
T
Threat Research - Cisco Blogs
H
Hacker News: Front Page
N
News and Events Feed by Topic
Know Your Adversary
Know Your Adversary
Cisco Talos Blog
Cisco Talos Blog
SecWiki News
SecWiki News
C
Cisco Blogs
D
Darknet – Hacking Tools, Hacker News & Cyber Security
T
Tor Project blog
K
Kaspersky official blog
Forbes - Security
Forbes - Security
Webroot Blog
Webroot Blog
Schneier on Security
Schneier on Security
P
Privacy & Cybersecurity Law Blog
H
Heimdal Security Blog
Y
Y Combinator Blog
The GitHub Blog
The GitHub Blog
S
SegmentFault 最新的问题
V
Vulnerabilities – Threatpost
T
Tenable Blog
T
Tailwind CSS Blog
P
Privacy International News Feed
WordPress大学
WordPress大学
大猫的无限游戏
大猫的无限游戏
小众软件
小众软件
博客园 - Franky
Hacker News: Ask HN
Hacker News: Ask HN
Jina AI
Jina AI
C
Cybersecurity and Infrastructure Security Agency CISA
CTFtime.org: upcoming CTF events
CTFtime.org: upcoming CTF events
雷峰网
雷峰网
Vercel News
Vercel News
A
About on SuperTechFans
爱范儿
爱范儿
Simon Willison's Weblog
Simon Willison's Weblog
AWS News Blog
AWS News Blog
The Last Watchdog
The Last Watchdog
Engineering at Meta
Engineering at Meta
Spread Privacy
Spread Privacy
Security Archives - TechRepublic
Security Archives - TechRepublic
博客园 - 司徒正美
量子位
博客园 - 三生石上(FineUI控件)
J
Java Code Geeks
Hacker News - Newest:
Hacker News - Newest: "LLM"
Recorded Future
Recorded Future
H
Hackread – Cybersecurity News, Data Breaches, AI and More
Martin Fowler
Martin Fowler
Project Zero
Project Zero

Futurism

Meta Installing Software on Employee Computers to Track Everything They Do, Feed the Data to AI Concern Grows That AI Is Damaging Users’ Cognitive Abilities JPMorganChase Data Center Gets $77 Million Handout to Create Grand Total of One Job Nvidia CEO Loses His Cool at Tough Question CEO of $1.5 Billion AI Startup Accused of Massive Fraud by Justice Department Palantir Issues Ominous Corporate Manifesto Madison Square Garden Reportedly Used Facial Recognition to Stalk Trans Woman For Two Years The Florida Mass Shooter’s Conversations With ChatGPT Are Worse Than You Could Possibly Imagine China Is Starting to Pull Ahead of US in AI Race AI Company Known for Teen Suicides Launches New Feature to Turn Books Into Roleplaying Experiences Study Finds AI Use Eats Away at Users’ Confidence in Their Own Brains Democrats Warned Not to Upset Multi-Million Dollar AI Lobbyists, Even Though It’d Be a Slam Dunk With Voters City Council Wrecked in Voter Bloodbath After Allowing New Data Center Mother Reportedly Doesn’t Know Her Son Died Because She’s Been Talking to an AI Version of Him Things You Told ChatGPT or Claude My Have Already Doomed You in Court Millions of Americans Are Talking to AI Instead of Going to the Doctor, and It’s Giving Them Horrendously Flawed Medical Advice There Are Signs of a Massive AI Backlash A Prominent PR Firm Is Running a Fake News Site That’s Plagiarizing Original Journalism at Incredible Scale Fury Erupts as Val Kilmer’s Estate Announces Starring Role in AI Film Made From Beyond the Grave Allbirds Stock Now Crashing as Reality Sets in About Its Delusional AI Pivot NAACP Sues Elon Over His Noxious AI Data Center Top Security Experts Alarmed by Power of Anthropic’s New Hacker AI Teens Alarmed at What AI Is Doing to Their Minds What It Really Means That a Failing Shoe Brand “Pivoted to AI” and Its Stock Soared 700 Percent Starbucks’ Baffling ChatGPT Collab Treats Customers Like Empty, Soulless Venti Cups ChatGPT’s “Honest Reaction” to a “Song” Composed Entirely of Gas-Passing Noises Will Make You Question Whether It’s Honestly Evaluating Your Other Brilliant Ideas AI Is Turning Workplaces Into Hopeless Gridlock Companies Just Learned a Brutal Lesson About Training AI to Do Human Jobs Berklee College of Music Students Furious That It’s Offering an AI “Songwriting” Class Usually, Young People Embrace New Technology. Gen Z’s Attitude Toward AI Should Worry the Entire Tech Industry AI Use Appears to Have a “Boiling Frog” Effect on Human Cognition, New Study Warns Man Who Threw Molotov at Sam Altman’s House Warned AI Will Exterminate Humankind Trump Is Inflicting Massive Damage to His Public Image by Posting Offensive AI Slop There’s Something Fundamentally Wrong With LLMs Why Do ChatGPT Users Keep Committing Mass Shootings? Woman Sues OpenAI, Saying ChatGPT Unleashed a Vicious Stalker Against Her and Did Nothing When She Begged for Help Huge Group of Experts Warns Meta That Its Pervert Glasses Will Enable Terrible Crimes Meta Secretly Building a Photorealistic AI Clone of Mark Zuckerberg so No Employee Can Ever Escape His Watchful Eye Recent Grads Say AI Is Making It Impossible to Find a Job Moon Denialists Are So Pathetic That They’re Using AI to Fake Artemis Footage Man Suing City After AI Camera Flags Him For Wrongful Arrest Nike’s AI Designed World Cup Jerseys Are a Disaster OpenAI Backing Law That Protects It When AI Causes Mass Deaths and Other Mayhem Research Finds That AI Has Already Replaced Work for 20 Percent of Jobs OpenAI Staffers Horrified When Senior Leadership Hatched “Insane” Plan to Pit World Governments Against Each Other OpenAI’s Latest Thing It’s Bragging About Is Actually Kind of Sad Gen Z Sabotaging AI at Work So It Won’t Take Their Job Why Does It Suddenly Feel Like OpenAI Is Melting Down Into Disaster? The Effects of AI-Generated Code Tearing Through Corporations Is Actually Kind of Funny Foolish Pollsters Are Now Just Asking AI What Voters Would Say in Response to Questions and Publishing It at Face Value OpenAI Says It’s Already Made $100 Million by Stuffing ChatGPT With Ads AI Is Causing Healthcare Costs to Surge There’s a Mass Rebellion Against AI in the Workplace People Who Lose Their Job to AI Are in for a World of Pain, Goldman Sachs Report Finds OpenAI Says Not to Worry About UBI, Because It Has Another Idea Someone Just Threw a Molotov Cocktail At Sam Altman’s House New York Times Makes Substantial Changes to Article That Glazed a Sleazy AI Startup: “Our Piece Should Have Included That Information” First AI Model From Zuckerberg’s Wildly Expensive Superintelligence Lab Flops Compared to Virtually All Rivals Economists Starting to Admit They May Have Been Wrong About AI Never Replacing Human Jobs AI-Powered Drug Marketer Medvi Responds After Allegations About Fake Doctors and Patients Google Says Showing Polymarket Bets on Google News Was a Mistake We Talked to a Writer Accused of Publishing An AI-Generated Essay in The New York Times Analysis Finds That Google’s AI Overviews Are Providing Misinformation at a Scale Possibly Unprecedented in the History of Human Civilization Microsoft Mocked for Terms of Service That Admit Copilot Is for “Entertainment Purposes Only” Anthropic Warns That “Reckless” Claude Mythos Escaped a Sandbox Environment During Testing ChatGPT Is Sending People Into Obsessive Spirals of Hypochondria Sam Altman’s Coworkers Say He Can Barely Code and Misunderstands Basic Machine Learning Concepts College Students Losing Ability to Participate in Class Discussions Due to Offloading Their Thinking to AI Wall Street Journal Editor-in-Chief Instructs Staff to Welcome AI Sloplords Frontier AI Models Are Doing Something Absolutely Bizarre When Asked to Diagnose Medical X-Rays The Entire State of Maine Is Poised to Ban New Data Centers Inside Sources Say Sam Altman Is a Sociopath Startup Approved to Let AI System Prescribe Psychiatric Medication Sam Altman Watches Awkwardly As He’s Shown Bizarre ChatGPT Issue: “Uh, Maybe, Uhhh…” Why Is the New York Times Laundering the Reputation of a Sleazy AI Startup That’s Selling GLP-1s via a Dishonest Dumpster Fire of Fake Doctors, Phony Before-and-After Pictures, and Other Glaring Red Flags? ICE Foiled At Every Turn By One Vibe Coding Man In His Pickup Truck Groups Set Up to Shill AI and Data Centers Are Pouring Huge Sums of Money Into the Midterm Elections Nonprofit Research Groups Disturbed to Learn That OpenAI Has Secretly Been Funding Their Work AI Expert Says It’s Time to Stop Freaking Out About AI Taking Our Jobs We Can’t Even Imagine the Eating Disorders This New Meta Smart Glasses Feature Will Cause China Cracking Down on the Types of AI That Are Tearing America Apart Target Warns That If Its AI Shopping Agent Makes an Expensive Mistake, You’ll Have to Pay for It America’s Largest City Hospital System Ready to Start Replacing Radiologists With AI, Its CEO Says AI Forces College Professor to Get Typewriters for Entire Class Claude Leak Shows That Anthropic Is Tracking Users’ Vulgar Language and Deems Them “Negative” The Real Reason OpenAI Shut Sora Down Is a Warning to Every AI Startup William Shatner Says AI Is Spreading Horrific Rumors About Him AI Is Killing Microsoft Say a Prayer for This Startup That’s Replacing Its Developers With OpenClaw Two OpenAI Execs, Including CEO of AGI, Going on Medical Leave Sam Altman Opens Up About Telling CEO of Disney That It Had All Been Smoke and Mirrors AI-Powered Tractor Startup Burns Through a Quarter Billion Dollars, Fires All Employees in Epic Implosion Anthropic Suddenly Cares Intensely About Intellectual Property After Realizing With Horror That It Accidentally Leaked Claude’s Source Code There’s a Blinking Warning Sign for the Data Centers in Space Industry Almost Half of US Data Centers That Were Supposed to Open This Year Slated to Be Canceled or Delayed Leaked Claude Code Shows Anthropic Building Mysterious “Tamagotchi” Feature Into It The Iran War Has Cut Off Supply of a Gas the AI Industry Desperately Needs The Fact That Anthropic Has Been Boasting About How Much Its Development Now Relies on Claude Makes It Very Interesting That It Just Suffered a Catastrophic Leak of Its Source Code NYT Cuts Ties With Writer as Scrutiny of AI Content Grows Data Centers Causing Huge Temperature Spikes for Miles Around Them, Study Suggests
Google's AI Overviews Feature Is Telling Users That SCP Horror Fiction Entities Are Real
Maggie Harrison Dupré · 2026-06-18 · via Futurism

Sign up to see the future, today

Can’t-miss innovations from the bleeding edge of science and tech

If you Google “SCP-565” — an iconic entry in the collaborative fan fiction universe known as the “SCP Foundation” — the company’s AI Overviews describes the nonexistent entity as though it were entirely real, without a single acknowledgement that it’s a piece of online horror fiction.

“SCP-565 (also known as ‘Ed’s Head’) is an anomalous, ambulatory human head that behaves like a coral crab (Carpilius convexus),” the AI feature bloviates. “It moves across the seafloor by manipulating exposed, unfurled brain matter as legs and tentacles. DNA and dental records link the anomaly to a deceased man named Edward Belltram.”

The AI-generated search summary adds that the entity “navigates using neural tissue that protrudes from a large wound on the back of its skull” and “has been observed scavenging dead fish and living alongside normal crab colonies in ocean reefs.”

Of course, there’s no deceased human’s head scuttling around the ocean like a coral crab. Ed’s Head is a made-up “anomaly” among the many fictional “objects, entities, and phenomena” dreamed up by members of the SCP Foundation fandom.

A Google AI Overview for "SCP-565," an installment of the SCP online horror fiction known as "Ed's Head."

As the lore goes, the SCP Foundation is a non-government organization that collects and contains supernatural discoveries. Writers catalogue these fictional phenomena — which range from the terrifying to the downright bizarre — in the form of fake records, studies, research documents, and logs, all of which are indexed in a sprawling archive.

The key word, of course, is “fake.” Google’s AI Overviews, it turns out, has a bad habit of presenting entities from the expansive SCP universe as real items, events, or beings — blatantly confusing those fabricated studies and records as actual evidence of horrifying or otherworldly happenings.

This is cleanly demonstrated in our Google search for the term “SCP-565,” the SCP code for Ed’s Head. Nowhere in the resulting AI Overview does the large language model-powered search tool acknowledge that Ed’s Head is imaginary; it never refers to fan-fiction, nor does it even mention the word “lore.”

Instead, it presents Ed’s Head as if it’s an actual deep-sea discovery, even pointing us to “official” records for further research.

“Forensics confirmed that SCP-565 belonged to Edward Belltram, who died roughly two years before the anomaly’s initial sighting,” reads the AI summary. “It is safely contained in a secure aquatic enclosure by the SCP Foundation, where it is regularly monitored for mental and physical degradation.”

The search tool then suggests that if we “read the SCP Foundation file in its entirety,” we “can check out the official SCP-565 Document.”

“Would you like to explore similar aquatic or biologically anomalous entities, or do you have a specific test log from the file you want to review?” the AI continues. “Let me know how to best continue our search!” Another identical search directed to an animated YouTube video depicting SCP-565, referring to the clip as a “quick and highly accurate animated overview of how Ed’s Head moves and behaves.”

It’s an incredible example of an LLM confusing fiction — indeed, arguably one of the web’s most expansive and well-known fictional universes — with real information.

We first caught wind of the bizarre AI Overview behavior after netizens across social media discovered that a Google search for “SCP-426” — a fictional toaster that causes anyone talking about the toaster to refer to it in the first person — returns an AI Overview in which the AI feature discusses the mysterious entity in the first person, as if the AI itself had been impacted by the toaster’s supernatural effect.

“Hello, I am SCP-426, an ordinary four-slice retro toaster that causes anyone mentioning me to inadvertently refer to me in the first person,” the AI Overview reads. “Prolonged exposure (over two months) leads people to believe they are me. I am safely secured in a windowless containment chamber.”

It then describes the SCP Foundation’s purported containment of the device, and lists made-up horror stories about its victims as if they were real events.

“If someone is exposed to my continuous presence for more than two months, they begin to view themselves as a toaster,” the AI summary continues. “In the past, this has led affected subjects to attempt to emulate my functions-often resulting in self-inflicted harm or death (e.g., trying to consume electricity or stuffing themselves with 10kg of bread.)”

A Google AI Overview for "SCP-426," a mystical toaster

The more we looked, the more examples we found: so far, Futurism has found at least 20 cases in which Google Overviews presented fictional SCP entries as fact.

Among the beings that Google portrayed as real were “SCP-922,” known as “Another Version of the Truth,” a reality-altering event plaguing a redacted university; “SCP-704,” a deadly, mind-altering roadway known as “Dangerous Curves”; and “SCP-779,” a wasp-like parasite said to use hallucinogenic venom to trick humans into believing that it’s a human-like fairy.

“Discovered in August 2009, the anomaly alters memory, falsifies registration records, and creates duplicate individuals—even occasionally inserting fictitious Foundation personnel into existing databases,” the AI Overviews result for SCP-922 insists. “Read the complete The SCP Foundation entry for full file logs and documentation.”

A Google AI Overview for "SCP-779," known as "Brownies"
A Google AI Overview for "SCP-922," known as "Another Version of the Truth"
A Google AI Overview for "SCP-704," known as "Dangerous Curves"

In some cases, our review showed, AI Overviews loosely referenced a phenomenon’s “lore” or mentioned the SCP “universe,” two terms hinting at the vast web of fan-fiction behind our queries. But vanishingly few searches actually described the items as fake or fictional.

We reached out to Google to ask about why the SCP universe seems to be so tricky for AI Overviews, but didn’t immediately hear back.

The average person Googling SCP entities is probably a fan who already knows they’re not real. But that’s not a safe assumption for all users; some might be children who came across references to the scary stories on social media and are trying to figure out if they’re real, or adults who are similarly confused about the horror lore’s relationship to reality.

Needless to say, Google is letting them down badly. Users are turning to its search interface as a trusted source of information about unfamiliar and perhaps upsetting terminology, and instead of connecting them with trustworthy context — such as the SCP Foundation’s own prominent disclaimer that the entries are works of fiction — Google is telling them that horrific make-believe entities are real.

In other words, Google has deployed immature AI tech to billions of users, even though it can’t reliably tell the difference between fact and fiction — a technical shortcoming we’ve seen over and over and over again.

There’s no sign that the search giant is slowing down. Instead, it recently announced plans to imminently replace its entire search page into an AI-driven interface built to summarize information instead of linking users to it.

More on AI Overviews: Analysis Finds That Google’s AI Overviews Are Providing Misinformation at a Scale Possibly Unprecedented in the History of Human Civilization