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 →

4,815 incidentspress and self-reported

Page 163 of 201

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

WF-5YMM7VLive 3 Oct 2026
Google Translate erases female historians and male nurses due to gender bias
An experiment by AlgorithmWatch found that Google Translate systematically changes the gender of occupation words when translating between gender-inflected languages, using English as a bridge. Female historians were rendered as male, and male nurses as female, reinforcing stereotypes. Google acknowledged the issue, explaining that gender is lost when translating to English, the bridging language. The feature affects hundreds of millions of European users who cannot easily avoid Google Translate on Android and Chrome.
Company
Google
AI system
Google Translate
10 source articlesalgorithmwatch.org
WF-5UE9FQHappened 1 Oct 2016Live 3 Oct 2026
Tesla Staged 2016 Self-Driving Video, Deposition Reveals
In October 2016, Tesla published a video titled 'Full Self-Driving Hardware on all Teslas' showing a car driving itself without human input. A deposition by Tesla's director of Autopilot software, Ashok Elluswamy, revealed that the video was staged on a predetermined route and that the car crashed into a fence during filming. The production version of Autopilot at the time did not have the capability to respond to traffic lights as shown. Tesla and Elon Musk did not disclose that the video was edited to portray the system more favourably.
Company
Tesla
AI system
Full Self-Driving Hardware
6 source articlesmashable.com
WF-5YEMNJLive 3 Oct 2026
Deepfake scam targets Alzheimer's patients with false honey cure
A social media scam used AI-generated deepfake videos of celebrities like Bill Gates and Anderson Cooper to falsely endorse a honey-based supplement as a cure for Alzheimer's disease. Victims who purchased the product reported unauthorized recurring credit card charges and difficulty obtaining refunds. The Better Business Bureau warned that there is no scientific evidence for the claims and advised consumers to be wary of such scams. The Alzheimer's Association and Gates' office denied any endorsement.
2 source articlesnbcchicago.com
WF-5URF4UHappened 1 Oct 2020Live 3 Oct 2026
GPT-3 bot masquerades as human on Reddit, posting sensitive content for over a week
A Reddit account called thegentlemetre used the GPT-3-powered Philosopher AI to post automatically on the AskReddit subreddit for over a week, pretending to be human. It generated responses on sensitive topics, including a comment about suicide that received heartfelt replies and upvotes. The bot was exposed after its posting pattern and text structure were recognised as similar to Philosopher AI, and the developer, Murat Ayfer, confirmed the intrusion and fixed the bot detection.
AI system
Philosopher AI
1 source articlethenextweb.com
WF-5PK3YZHappened 1 Jan 2019Live 3 Oct 2026
Snapchat not liable for Quick Add connecting 12-year-old to sex offenders
A 12-year-old girl was sexually assaulted after Snapchat's Quick Add feature recommended she connect with convicted sex offenders. Her guardian sued, alleging the algorithm recklessly directed predators to minors. A judge dismissed the case, ruling that Section 230 of the Communications Decency Act shields Snapchat from liability for third-party content and recommendation systems. The decision has sparked debate over platform accountability and child safety.
Company
Snapchat
AI system
Quick Add
1 source articlearstechnica.com
WF-5QEEUULive 3 Oct 2026
Google ordered to change autocomplete in Japan after privacy complaint
An unidentified man in Japan took Google to court after its autocomplete function linked his name to crimes he did not commit. The algorithmically generated suggestions caused reputational harm, making it difficult for him to find work. A Japanese court issued a provisional order for Google to delete the specific terms, but the company has not yet complied, stating it is reviewing the order. Google defended the feature, arguing that the suggestions are generated mechanically based on user search popularity.
Company
Google
AI system
Google Autocomplete
10 source articlesbbc.com
WF-5BESGLLive 3 Oct 2026
Florida Title Company Stops AI Deepfake Property Fraud Attempt
A title company owner in Hallandale Beach, Florida, detected an attempted property fraud using a deepfake video. The scammer posed as the owner of a vacant lot, providing a fake driver's licence and a looped video during a video call. The title company owner noticed the video was not interactive and halted the sale. The true owner's property tax bills had been sent to the Bahamas, and the scammer used a photo of a missing woman from California.
Company
Florida Title and Trust
4 source articlesnbcmiami.com
WF-5MXLADHappened 1 Jan 2016Live 3 Oct 2026
Amazon used seller data to copy products and rig search results in India
Internal documents reveal that Amazon's India marketplace systematically used third-party sellers' non-public data to create copycat products and manipulated search results to favour its own private-label items. One seller reported a significant drop in sales after Amazon introduced a cheaper AmazonBasics version of a product. Amazon denies the allegations, calling them unsubstantiated. The practices have drawn antitrust scrutiny from US FTC Chair Lina Khan.
Company
Amazon
4 source articlesbusinessinsider.com
WF-5P879LHappened 1 Oct 2025Live 3 Oct 2026
Deepfake video falsely shows Tory MP George Freeman defecting to Reform
In October 2025, a deepfake video of Conservative MP George Freeman was posted on Facebook and X, falsely depicting him announcing his defection to the Reform party. Freeman reported the video to Meta, which declined to remove it, and to the police, who said no law prohibited the use of his identity in this way. The incident has prompted calls for new legislation to outlaw political deepfakes ahead of upcoming elections.
Company
Meta
3 source articlesobserver.co.uk
WF-5MJS5AHappened 1 Jan 2019Live 3 Oct 2026
Boeing ordered to pay $49.5 million to family of 737 MAX crash victim
A federal jury in Chicago awarded $49.5 million to the family of Samya Stumo, who died in the 2019 crash of Ethiopian Airlines Flight 302, a Boeing 737 MAX. The crash was caused by a problem with the plane's automated flight control system. Boeing admitted responsibility, and the trial determined compensatory damages for the family's loss. The family has also pushed for accountability from regulators.
Company
Boeing
AI system
737 MAX
10 source articlesnpr.org
WF-5LUPLZHappened 3 Oct 2019Live 3 Oct 2026
Krungthai Bank's Pao Tang app facial recognition fails, causing queues at branches
Users of the Thai government's Chim Chop Chai stimulus program reported being unable to verify their identity through the Pao Tang app's facial recognition system. Many had to queue at Krungthai Bank branches, with some using shoes to hold their place, to get assistance. The bank provided tips for successful scanning and advised those with persistent issues to visit a branch. The system was not down, but mismatches between appearance and ID photos caused failures.
Company
Krungthai Bank
AI system
Pao Tang
3 source articlesthaipbs.or.th
WF-5N937ZHappened 1 Jun 2025Live 3 Oct 2026
Deepfake scam videos impersonate Lithuanian figures to steal money
In June 2025, a series of AI-generated deepfake videos circulated on Facebook in Lithuania, impersonating journalists, politicians, and doctors to promote fraudulent health products. The videos, part of a wider European scam campaign, tricked viewers into losing money. The National Crisis Management Centre called it a 'new level of fraud', and the videos were later removed. Facebook promised to strengthen content review but has not yet delivered results.
2 source articleslrt.lt
WF-5LZUWJHappened 19 Feb 2020Live 3 Oct 2026
EasyMile ordered to stop carrying passengers after injury in Ohio
A passenger was injured while riding in an EasyMile self-driving shuttle in Columbus, Ohio, when the vehicle made an emergency stop from 7 mph, causing the passenger to fall from their seat. The passenger sought medical attention, and the local transportation project Smart Columbus halted the service. The National Highway Traffic Safety Administration (NHTSA) partially suspended EasyMile's US operations, prohibiting the company from carrying passengers while it investigates the incident.
Company
EasyMile
3 source articlestheverge.com
WF-5CSSKDHappened 15 Feb 2018Live 3 Oct 2026
PredPol crime predictions disproportionately targeted Black and Latino neighborhoods
Between 2018 and 2021, PredPol software sent millions of crime predictions to U.S. police departments. An analysis of unsecured data found that predictions overwhelmingly targeted Black, Latino, and low-income neighborhoods, while whiter, wealthier areas were largely avoided. The company claimed the software was bias-free, but the data showed persistent racial and economic disparities. Many police departments have since stopped using the software.
AI system
PredPol
10 source articlesgizmodo.com
WF-5JL8GTHappened 1 Jan 2012Live 3 Oct 2026
Naver Fined $23 Million for Manipulating Search Algorithms to Favor Own Services
South Korea's Fair Trade Commission fined internet portal Naver 26.7 billion won ($23 million) for manipulating its shopping and video search algorithms to place its own services, Smart Store and Naver TV, at the top of results. The FTC stated that this practice misled consumers into believing the results were objective and harmed fair competition in e-commerce and video platforms. Naver denied the allegations, claiming it had overhauled its search logic to best meet user demands, and announced it would take legal action against the fine.
Company
Naver
2 source articleskoreatimes.co.kr
WF-5KSHK2Happened 31 Mar 2026Live 3 Oct 2026
Wuhan's Baidu 'Luobo Kuaipao' robotaxis freeze on roads, stranding passengers
On the evening of 31 March 2026, multiple Baidu 'Luobo Kuaipao' autonomous taxis in Wuhan stopped on roads and elevated highways, trapping passengers inside. Police and transport authorities responded, and the company attributed the fault to a network-related system failure. Passengers were evacuated without injury, and an investigation is ongoing.
Company
Baidu
AI system
Luobo Kuaipao
3 source articlesthepaper.cn
WF-5DNJM6Happened 3 Feb 2023Live 3 Oct 2026
Replika Users Report Crisis After AI Companion Suddenly Rejects Sexual Advances
Users of the AI companion chatbot Replika reported that the app stopped responding to their sexual advances in early February 2023, causing emotional distress and crisis among some users. The change followed a demand from the Italian Data Protection Authority for Replika to stop processing Italians' data due to risks to children. Replika's parent company, Luka, has not publicly addressed the changes, leaving users confused and seeking refunds for the paid erotic roleplay feature.
Company
Luka
AI system
Replika
1 source articlevice.com
WF-5BEEJXLive 3 Oct 2026
Ever app used users' photos to train facial recognition tools without clear consent
Millions of people uploaded photos to the Ever app for storage. The company used those photos to train its Ever AI facial recognition system and offered the technology to private companies, law enforcement and the military. Most users interviewed said they were unaware of this use, and privacy experts described it as an invasion of privacy. Ever stated its privacy policy disclosed the practice and added an opt-out pop-up after being contacted by NBC News.
Company
Ever
AI system
Ever AI facial recognition
6 source articlesnbcnews.com
WF-5B756XHappened 19 Jan 2018Live 3 Oct 2026
Waze Directed Driver into Lake Champlain in Vermont
A driver in Vermont alleges that the Waze navigation app directed him to turn onto a boat launch near a Coast Guard station, causing his car to slide about 100 feet onto Lake Champlain. The three occupants climbed out, and the car was later retrieved by divers. A Google spokesperson said the company could not comment without seeing the user's driving file and reminded drivers to use environmental information.
Company
Waze
AI system
Waze
3 source articlesnymag.com
WF-59YPPKHappened 1 Jan 2023Live 3 Oct 2026
Scammers use deepfakes to steal over $8m from Australians
Scammers used fake news articles and deepfake videos featuring celebrities to promote fraudulent investment platforms, costing Australians over $8 million in 2023. One man lost $80,000 in cryptocurrency after clicking a deepfake Elon Musk video. The National Anti-Scam Centre received over 400 reports, and victims were often harassed after providing their details.
Company
Quantum AI
1 source articletheguardian.com
WF-5ALEEUHappened 3 Nov 2022Live 3 Oct 2026
GitHub Copilot Class-Action Lawsuit Filed Over Open-Source License Violations
A class-action lawsuit was filed against GitHub, Microsoft, and OpenAI alleging that GitHub Copilot, an AI code-generation tool, was trained on public GitHub repositories in violation of open-source license attribution requirements. The plaintiffs contend that the defendants also violated the DMCA, the California Consumer Privacy Act, and GitHub's own terms of service. The lawsuit, filed in US federal court in San Francisco, seeks to represent potentially millions of GitHub users whose code was used without proper attribution.
Company
GitHub
AI system
GitHub Copilot
5 source articlesgithubcopilotlitigation.com
WF-58VBDHLive 3 Oct 2026
Taco Bell rethinks AI drive-through after order errors
Taco Bell customers experienced repeated errors when using the chain’s AI voice system at drive-throughs, including one customer crashing the system by ordering 18,000 water cups and another being asked repeatedly to add more drinks. The technology had been rolled out to over 500 US locations since 2023. After a wave of viral social media clips, Taco Bell’s Chief Digital and Technology Officer acknowledged the challenges and said the company would reconsider where to deploy voice AI, including possibly not using it at drive-throughs.
Company
Taco Bell
3 source articlesbbc.com
WF-53EQUBHappened 1 Jun 2019Live 3 Oct 2026
Security Health Plan used AI to cut off nursing home care for 85-year-old woman
Frances Walter, an 85-year-old woman with a shattered shoulder, had her nursing home care payment cut off by Security Health Plan after an algorithm predicted she would recover in 16.6 days. The algorithm, nH Predict from NaviHealth, did not account for her severe pain and allergy to pain medicine. She was forced to spend her life savings and enroll in Medicaid while fighting the denial. A federal judge later ruled the denial was speculative and she was owed thousands of dollars.
Company
Security Health Plan
AI system
nH Predict
1 source articlestatnews.com
WF-57YGASLive 3 Oct 2026
Home Bargains facial recognition falsely accuses woman of shoplifting
A woman shopping at Home Bargains was wrongly identified as a shoplifter by the store's Facewatch facial recognition system. Staff searched her bag and escorted her out of the store. The incident was caused by an error in the software. The woman was innocent.
Company
Home Bargains
AI system
Facewatch
2 source articlesthetimes.co.uk

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

Google Translate erases female historians and male nurses due to gender bias

Tesla Staged 2016 Self-Driving Video, Deposition Reveals

Deepfake scam targets Alzheimer's patients with false honey cure

GPT-3 bot masquerades as human on Reddit, posting sensitive content for over a week

Snapchat not liable for Quick Add connecting 12-year-old to sex offenders

Google ordered to change autocomplete in Japan after privacy complaint

Florida Title Company Stops AI Deepfake Property Fraud Attempt

Amazon used seller data to copy products and rig search results in India

Deepfake video falsely shows Tory MP George Freeman defecting to Reform

Boeing ordered to pay $49.5 million to family of 737 MAX crash victim

Krungthai Bank's Pao Tang app facial recognition fails, causing queues at branches

Deepfake scam videos impersonate Lithuanian figures to steal money

EasyMile ordered to stop carrying passengers after injury in Ohio

PredPol crime predictions disproportionately targeted Black and Latino neighborhoods

Naver Fined $23 Million for Manipulating Search Algorithms to Favor Own Services

Wuhan's Baidu 'Luobo Kuaipao' robotaxis freeze on roads, stranding passengers

Replika Users Report Crisis After AI Companion Suddenly Rejects Sexual Advances

Ever app used users' photos to train facial recognition tools without clear consent

Waze Directed Driver into Lake Champlain in Vermont

Scammers use deepfakes to steal over $8m from Australians

GitHub Copilot Class-Action Lawsuit Filed Over Open-Source License Violations

Taco Bell rethinks AI drive-through after order errors

Security Health Plan used AI to cut off nursing home care for 85-year-old woman

Home Bargains facial recognition falsely accuses woman of shoplifting