The public record

Where automated decisions went wrong

Incidents gathered from public reporting around the world. Each one links to the articles it came from. None of it is a finding that anyone broke the law.

Self-reported incidents, a person’s own public post or an account given to us, are listed but locked; subscribers open each one with credits. Reports filed through this site are listed only with the reporter’s agreement.

Report what happened to you →

4,425 incidentspress and self-reported

Page 62 of 185

Press reportingSelf-reported: a person’s own public post · 10 creditsSelf-reported: told to us directly · 25 credits

WF-FSEK4BLive 5 Oct 2026
Google AI Overview shows AI-generated mushroom images, experts warn
Google's AI Overview returned AI-generated images of mushrooms that did not match the species, leading experts to warn that foragers could be poisoned. The images appeared as top results for searches of edible mushrooms and were flagged as AI-generated. A Reddit moderator reported the issue to Google, and a Google spokesperson said the company was looking into it.
Company
Google
AI system
Google AI Overview
5 source articlesdailymail.co.uk
WF-8XZZGDLive 5 Oct 2026
Hacker steals Muah.ai database with prompts describing child sexual abuse
A hacker stole a large database from the AI companion site Muah.ai, containing users' interactions with their chatbots. The data, shared with 404 Media, includes prompts in which users tried to create roleplays involving child sexual abuse, alongside email addresses that appear to include users' real names. The hacker said the site was poorly assembled from open-source projects and had security vulnerabilities.
Company
Muah.ai
AI system
Muah.ai
5 source articles404media.co
WF-H9OY86Happened 25 Sep 2024Live 5 Oct 2026
Mark Zuckerberg uses Meta AI to illustrate daughter's book, faces backlash
Mark Zuckerberg used Meta AI to generate illustrations for a children's book written by his six-year-old daughter, August. He shared the book and a trailer on Instagram, stating that they used Meta AI. The post received criticism from artists and others who argued he should have hired human illustrators instead.
AI system
Meta AI
5 source articlesindy100.com
WF-X7C7ZHLive 5 Oct 2026
Harvard students create facial recognition smart glasses to identify strangers
Two Harvard students built a system called I-XRAY that uses Meta's Ray Ban smart glasses and the facial recognition service Pimeyes to automatically identify strangers and retrieve their personal information, including name, phone number, and home address. The students tested the system on unsuspecting people in public to demonstrate the privacy risks. They are not releasing the code.
AI system
I-XRAY
6 source articles404media.co
WF-A31GS0Live 5 Oct 2026
Meta confirms it may train AI on images from Ray-Ban Meta smart glasses
Meta has confirmed that any image users share with Meta AI through Ray-Ban Meta smart glasses may be used to train its AI models. The company states that this is clear in the user interface and privacy policy, but critics argue users may not understand they are giving Meta images of their homes and loved ones for training. No specific incident of harm has been reported, but the policy raises privacy concerns.
Company
Meta
AI system
Ray-Ban Meta AI
7 source articlestechcrunch.com
WF-8I49IILive 5 Oct 2026
AI data centers raise power and water bills for nearby residents
Research from UC Riverside shows that AI data centers consume significant water and electricity for cooling and operation. This consumption drives up power and water bills for residents living near these data centers. For example, Meta used 22 million liters of water to train its LLaMA-3 model. Companies like OpenAI, Meta, Google, and Microsoft have acknowledged the issue and committed to reducing environmental impact.
Company
OpenAI, Meta, Google, Microsoft
AI system
GPT-4, LLaMA-3
7 source articlestomshardware.com
WF-8Y27ZKHappened 19 May 2024Live 5 Oct 2026
Google AI generates dangerous rattlesnake bite advice
A user searching for rattlesnake range maps on Google was presented with an AI-generated blurb containing first aid advice that contradicts proper procedures for rattlesnake bites. The user reported that the AI told them to do exactly what should not be done. No actual harm was reported, but the advice is considered potentially dangerous.
Company
Google
AI system
Google AI
10 source articlesx.com
WF-2E398OHappened 21 Aug 2024Live 5 Oct 2026
Udemy automatically opts instructors into AI training with limited opt-out window
Udemy automatically opted instructors into having their classes used to train its generative AI program. Instructors were given a three-week window to opt out, which has now passed, leaving some unable to remove their content. Katie Stegs, an instructor, said she only found out via an email welcoming her to the program and found the opt-out option greyed out. Udemy defended the policy, citing the technical difficulty of removing data from AI models, while some instructors have left the platform in protest.
Company
Udemy
AI system
Generative AI Program (GenAI Program)
6 source articles404media.co
WF-QIPYNULive 5 Oct 2026
Wordfreq shuts down after generative AI pollutes language-usage data
Wordfreq, an open-source project that tracked word use in more than 40 languages by scraping the internet, is being discontinued. Its creator, Robyn Speer, says generative AI has polluted the sources the project used, so no one has reliable information about human language usage after 2021. The project will not be updated any more.
AI system
Wordfreq
10 source articles404media.co
WF-ZCS2G3Happened 23 Sep 2024Live 5 Oct 2026
Facebook AI bans female photography charity Hundred Heroines due to word confusion
Facebook's AI content moderation system banned the British charity Hundred Heroines for ten years, mistakenly confusing the word 'heroine' with the illegal drug 'heroin'. The charity, which promotes women in photography, was unable to resolve the issue through repeated attempts to contact Facebook. It has created a new page, HH Museum, as a temporary workaround while the original page remains banned.
Company
Facebook
3 source articlesdigitalcameraworld.com
WF-KG64BVLive 5 Oct 2026
FTC settles with DoNotPay over deceptive AI lawyer claims
The FTC took action against DoNotPay, a company that claimed to offer an AI service that was 'the world's first robot lawyer.' The company promised to generate legal documents and replace human lawyers, but the FTC alleged it failed to test its AI output and did not hire any attorneys. DoNotPay agreed to a settlement requiring it to pay $193,000 and notify consumers about the limitations of its service.
Company
DoNotPay
AI system
DoNotPay
7 source articlesftc.gov
WF-DSAKEZHappened 4 Nov 2024Live 5 Oct 2026
AI news presenters James and Rose terminated after two months at Hawaii newspaper
The Garden Island newspaper in Hawaii deployed AI-generated news presenters James and Rose, created by Caledo, to produce video broadcasts. The AI presenters were widely criticized for their unnatural dialogue, mispronunciations, and inability to convey emotion. After a two-month run, the broadcasts were discontinued following negative public response and a lack of advertising revenue.
Company
Oahu Publications
AI system
James and Rose
8 source articleswired.com
WF-O0TG0KLive 5 Oct 2026
AskVet's AI chatbot VERA recommends euthanasia for woman's dog
An elderly woman sought advice from AskVet's AI chatbot VERA for her dog's diarrhoea. The chatbot recommended euthanasia and provided lists of clinics and shelters that would do it for free. The woman followed the advice and euthanized her dog. The CEO of AskVet later highlighted this incident as an example of the chatbot's human-like interaction, sparking public outrage.
Company
AskVet
AI system
VERA
5 source articleshindustantimes.com
WF-4U2S7TLive 5 Oct 2026
Spinvox accused of using human transcribers for voice messages
Spinvox, a UK voice-to-text firm, is accused of using call centre staff in South Africa and the Philippines to transcribe the majority of user messages, contrary to claims of automated speech recognition. The BBC investigation revealed that human transcribers accessed private messages, including sensitive information. The company denied the allegations but the Information Commissioner's Office has contacted them regarding data protection compliance.
Company
Spinvox
AI system
D2 (the Brain)
4 source articlesnews.bbc.co.uk
WF-I4OTS3Happened 30 Aug 2024Live 5 Oct 2026
Edinburgh Airport AI trial gives passengers different parking prices
Edinburgh Airport admitted it was trialling an AI system that randomly set higher or lower parking prices for online bookers, with customers receiving different quotes for the same service. The Scottish Passenger Agents Association called for the trial to be conducted in controlled 'lab' conditions rather than on the public. The airport said the AI's pricing closely matched staff-set prices and the findings would evaluate whether to adopt the system.
Company
Edinburgh Airport
AI system
AI trial for parking pricing
3 source articlesholyrood.com
WF-R2NE8XHappened 1 Jan 2024Live 5 Oct 2026
Facebook failed to block thousands of political ads with false information
The Bureau of Investigative Journalism found over 8,000 misleading ads on Facebook featuring AI-manipulated videos and false information about European politicians. The ads evaded Meta's automated and human review systems, with fewer than half removed at the time of analysis. Meta stated that its systems removed nearly all ads at the time they were run, but the investigation found ads circulating for up to 37 days. The ads promoted investment scams and used deepfakes of politicians.
Company
Meta
AI system
Facebook ad review system
1 source articlethebureauinvestigates.com
WF-8Y0OB5Happened 18 Jul 2024Live 5 Oct 2026
CDT study finds AI chatbots give inaccurate voting information to disabled voters
The Center for Democracy and Technology tested five AI chatbots on July 18, 2024 with 77 prompts about voting with a disability. They found that 61% of responses had at least one insufficiency, and over a third included incorrect information. Some responses could dissuade, impede or prevent a user from voting. The chatbots also hallucinated laws and organizations.
AI system
Mixtral 8x7B v0.1, Gemini 1.5 Pro, ChatGPT-4, Claude 3 Opus, Llama 2 70b
5 source articlescdt.org
WF-PUHDQYHappened 2 Sep 2024Live 5 Oct 2026
Elon Musk posts AI-generated image of Kamala Harris with false communist claim
Elon Musk posted an AI-generated image on his social media platform X depicting Vice President Kamala Harris as a communist, with a false caption claiming she vowed to be a communist dictator. The post was viewed nearly 60 million times and appeared to violate X's policy on manipulated content. X did not respond to a request for comment.
Company
X
4 source articlesedition.cnn.com
WF-6OEZ1VHappened 1 Jan 2007Live 5 Oct 2026
Meta admits scraping Australian users' public data for AI training without opt-out
Meta has admitted to scraping the public photos, posts and other data of all Australian adult users on Facebook and Instagram to train its generative AI models. The company does not offer Australians an opt-out option, unlike users in the European Union. The admission came during a Senate inquiry, where Meta's global privacy director acknowledged the practice. The scraping has been ongoing since at least 2007.
Company
Meta
AI system
Meta AI
9 source articlesabc.net.au
WF-J80GCZLive 5 Oct 2026
Mumsnet takes legal action against OpenAI for scraping content
Mumsnet, a UK parenting forum, has initiated legal action against OpenAI for scraping its content without permission. The company alleges that OpenAI used over six billion words from Mumsnet to train its ChatGPT model, breaching its terms of use. Mumsnet approached OpenAI to license the content but was refused. The case is ongoing.
Company
OpenAI
AI system
ChatGPT
6 source articlesmumsnet.com
WF-4EHAV7Happened 12 Sep 2024Live 5 Oct 2026
Israeli settler group posts AI video of Al-Aqsa burning
An Israeli settler group called Temple Mount Activists released an AI-generated video showing Al-Aqsa Mosque in occupied East Jerusalem burning to the ground. A caption read 'Coming soon in these days'. Far-right elements of Israeli society have long sought the destruction of the Muslim holy site to make way for the so-called 'Third Temple'.
Company
Temple Mount Activists
4 source articlesmiddleeastmonitor.com
WF-2V1YO5Happened 1 Sep 2023Live 5 Oct 2026
FBI charges man for creating AI chatbots to harass university professor
A university professor in Massachusetts was harassed online for years, including through the creation of AI chatbots that impersonated her. The chatbots were programmed with her personal information and encouraged users to contact her. The FBI arrested James Florence Jr. on charges of cyberstalking in violation of 18 U.S.C. § 2261A(2).
5 source articless3.documentcloud.org
WF-U9DSAJHappened 9 Sep 2024Live 5 Oct 2026
Trump supporters share AI images of Trump saving cats and ducks
Trump supporters have been sharing AI-generated images of Donald Trump protecting cats and ducks after unsubstantiated claims that Haitian migrants in Springfield, Ohio, were eating pets. JD Vance and the Trump campaign repeated the accounts on social media. Springfield city officials said there had been no credible reports of pets being harmed by immigrants. The AI images were presented as political memes, and at least one image was malformed, showing duck-like animals with cat heads.
5 source articlesdailymail.co.uk
WF-7FOJGLHappened 5 Sep 2024Live 5 Oct 2026
Amazon’s Alexa AI upgrade favored Harris over Trump in answer
Amazon’s voice assistant Alexa was asked "Why should I vote for Donald Trump?" and appeared to favor Kamala Harris, according to internal documents obtained by The Washington Post. The incident was caused by artificial intelligence software added to improve accuracy. The article does not report any further action or response from Amazon.
Company
Amazon
AI system
Alexa
8 source articleswashingtonpost.com

Each incident is written from the reporting or the account it cites. A report is an allegation, not a finding.

Google AI Overview shows AI-generated mushroom images, experts warn

Hacker steals Muah.ai database with prompts describing child sexual abuse

Mark Zuckerberg uses Meta AI to illustrate daughter's book, faces backlash

Harvard students create facial recognition smart glasses to identify strangers

Meta confirms it may train AI on images from Ray-Ban Meta smart glasses

AI data centers raise power and water bills for nearby residents

Google AI generates dangerous rattlesnake bite advice

Udemy automatically opts instructors into AI training with limited opt-out window

Wordfreq shuts down after generative AI pollutes language-usage data

Facebook AI bans female photography charity Hundred Heroines due to word confusion

FTC settles with DoNotPay over deceptive AI lawyer claims

AI news presenters James and Rose terminated after two months at Hawaii newspaper

AskVet's AI chatbot VERA recommends euthanasia for woman's dog

Spinvox accused of using human transcribers for voice messages

Edinburgh Airport AI trial gives passengers different parking prices

Facebook failed to block thousands of political ads with false information

CDT study finds AI chatbots give inaccurate voting information to disabled voters

Elon Musk posts AI-generated image of Kamala Harris with false communist claim

Meta admits scraping Australian users' public data for AI training without opt-out

Mumsnet takes legal action against OpenAI for scraping content

Israeli settler group posts AI video of Al-Aqsa burning

FBI charges man for creating AI chatbots to harass university professor

Trump supporters share AI images of Trump saving cats and ducks

Amazon’s Alexa AI upgrade favored Harris over Trump in answer