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 →

4,459 incidentspress and self-reported

Page 94 of 186

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

WF-UNLAI0Happened 1 May 2021Live 5 Oct 2026
Twitter's image-cropping algorithm found to favor young, thin females
Twitter's image-cropping algorithm was found to favor younger, thinner, and female faces over others, as well as white faces and Western text. The bias was discovered through a contest at the Defcon security conference where researchers analyzed the algorithm's code. Twitter acknowledged the bias and stopped using the algorithm in May 2021. The findings highlight the need for external scrutiny of AI systems to detect harmful biases.
Company
Twitter
10 source articleswired.com
WF-ANSGFVHappened 1 Dec 2022Live 5 Oct 2026
Apple kills plan to detect CSAM in iCloud
Apple announced a plan to detect child sexual abuse material in iCloud in 2021, but paused and then killed the plan in December 2022. Child safety advocates, including Sarah Gardner, are now engaging with Apple to urge implementation of the technology. The letter expresses concern about the delay and the impact on survivors.
Company
Apple
10 source articlesdocumentcloud.org
WF-FSZ9XQHappened 15 Apr 2021Live 5 Oct 2026
Apple Watch algorithm changes raise research data concerns
Harvard biostatistician JP Onnela found that heart rate variability data exported from an Apple Watch at two different times showed inconsistencies for the same time period, suggesting Apple's algorithms change without warning. The discovery led Onnela's team to abandon plans to use Apple Watches in research studies. Apple said changes to its algorithms only apply to data going forward and did not explain the discrepancy.
Company
Apple
AI system
Apple Watch
6 source articlestheverge.com
WF-Y5KMLCHappened 1 May 2021Live 5 Oct 2026
Social media platforms fail to act on 84% of antisemitic posts
A report by the Center for Countering Digital Hate found that social media platforms Facebook, Instagram, TikTok, Twitter, and YouTube failed to act on 84% of antisemitic posts reported by users. The posts, collected between May and June 2021, included Holocaust denial, conspiracy theories, and racist caricatures. The platforms did not remove the content despite their own community guidelines.
Company
Facebook, Instagram, TikTok, Twitter, YouTube
9 source articlescounterhate.com
WF-IK4RSLHappened 1 Jul 2016Live 5 Oct 2026
CBSA tested facial recognition on millions at Toronto Pearson Airport in 2016
In 2016, the Canada Border Services Agency (CBSA) secretly tested facial recognition technology on millions of travellers at Toronto Pearson Airport without their knowledge. The system, supplied by Face4 Systems, matched faces against a database of previously deported individuals. The CBSA stated that no one was deported as a result of the pilot, and the technology was removed after six months. Privacy advocates raised concerns about lack of consent, transparency, and potential racial bias.
Company
Canada Border Services Agency
AI system
Faces on the Move
8 source articlestheglobeandmail.com
WF-DA2WQJHappened 4 Feb 2021Live 5 Oct 2026
False AIS tracks simulated for Swedish Navy vessels
An investigation by SkyTruth and Global Fishing Watch revealed that false automatic identification system (AIS) tracks were generated for nine Swedish Navy vessels on February 4-5, 2021. The Swedish Navy confirmed the positions were false. The false tracks appeared plausible and were indistinguishable from real AIS broadcasts, posing a threat to data integrity for marine monitoring.
AI system
Automatic Identification System (AIS)
10 source articlesskytruth.org
WF-WM35TWHappened 1 May 2021Live 5 Oct 2026
Mercadona fined €2.5 million for facial recognition pilot
Mercadona, a Spanish supermarket chain, tested an early-detection system using facial recognition in 48 stores to identify people with judicial restraining orders. The system, which operated with court authorisation, notified police when such a person was detected. The Spanish data protection authority (AEPD) imposed a €2.5 million fine, which Mercadona paid, and the company has since removed the system citing legal uncertainty.
Company
Mercadona
AI system
Sistema de Detección Anticipada (Early Detection System)
10 source articleselpais.com
WF-ATTCQXLive 5 Oct 2026
Virginia crime lab faces first challenge to secret DNA algorithm
A defendant in an armed-robbery case in Fairfax County is challenging a secret algorithm used by the Virginia crime lab to interpret DNA evidence. The lab could not make a conventional match because skin-cell DNA on the victim's shirt was mixed with DNA from too many people; the algorithm identified the defendant as a contributor. The defendant is seeking to scrutinise the algorithm, reportedly for the first time.
Company
Virginia crime lab
10 source articleswashingtonpost.com
WF-LBFMV8Happened 12 Dec 2014Live 5 Oct 2026
Amazon sellers lose thousands after RepricerExpress pricing glitch
A software glitch in RepricerExpress's automated repricing system caused products sold by Amazon sellers to be priced at 1p, leading to significant financial losses for the sellers. RepricerExpress apologised but did not offer compensation, while Amazon said it had reached out to affected sellers but many reported no response. Some sellers have instructed lawyers to take legal action against RepricerExpress and Amazon.
Company
RepricerExpress
AI system
RepricerExpress
10 source articlestheguardian.com
WF-EEPCVNHappened 25 Jul 2021Live 5 Oct 2026
UAE cloud seeding drones cause Dubai rainstorms
The UAE government used cloud seeding drones to generate rainstorms in Dubai during a heatwave. The drones, developed with UK researchers, delivered electrical shocks to clouds to encourage precipitation. The resulting downpour was acknowledged by authorities as a deliberate weather modification effort.
Company
United Arab Emirates government
AI system
Cloud seeding drones
10 source articlesfuturism.com
WF-4TDRZRHappened 1 Jan 2019Live 5 Oct 2026
Doctored audio evidence used to discredit Dubai father in UK custody case
A woman used software to create a doctored audio recording of her husband, a Dubai resident, in which he appeared to make violent threats, and presented it as evidence in a UK child custody case. The court dismissed the recording after experts found it had been manipulated, but the case remains ongoing. The father's lawyer said the attempt was to deny him access to his children and portray him as dangerous.
6 source articlesthenationalnews.com
WF-OTMRHJHappened 24 Jun 2021Live 5 Oct 2026
Twitter Japan suspends accounts of government critics
Twitter Japan suspended and then restored the accounts of several critics of the Japanese government, including an account promoting a satirical film about Prime Minister Suga. The suspensions appear to be automated and were enacted without explanation, but were reversed after public outcry. The affected users allege that Twitter's moderation is biased against Japanese-language critics.
Company
Twitter Japan
8 source articlesglobalvoices.org
WF-M3S94THappened 18 Jan 2020Live 5 Oct 2026
Facebook apologizes after vulgar translation of Chinese leader's name
Facebook's translation system produced vulgar translations of a headline about Chinese President Xi Jinping and a post by Myanmar State Secretary Aung San Suu Kyi. The translations appeared on Facebook's platform, causing embarrassment. Facebook apologized for the errors.
Company
Facebook
AI system
Facebook translation
10 source articlesjpost.com
WF-YF44K5Live 5 Oct 2026
Pepper robot has severe security vulnerabilities allowing remote takeover
Researchers from Örebro University and Technical University of Denmark discovered multiple security flaws in SoftBank's Pepper robot, including unencrypted password transmission and lack of authentication for control commands. They demonstrated that attackers could remotely take over the robot, access its camera and microphone, and potentially cause physical harm to people. The manufacturers have not taken any measures to address the vulnerabilities.
Company
SoftBank Mobile and Aldebaran Robotics
AI system
Pepper
9 source articlesheise.de
WF-GUKKC3Happened 12 Oct 2019Live 5 Oct 2026
Tapia robot vulnerability at Henn na Hotels allows remote spying on guests.
A security researcher disclosed a vulnerability in the Tapia robot deployed at Henn na Hotels (Robot Hotels) in Japan. The robot accepts unsigned code via NFC, allowing an attacker to gain remote access to its camera and microphone. The researcher reported the issue to the vendor 90 days prior but received no response. The vulnerability potentially affects all future hotel guests.
Company
Henn na Hotels
AI system
Tapia robot
10 source articlestwitter.com
WF-ADQTAWLive 5 Oct 2026
Robot priests in Japan conduct Buddhist ceremonies
According to a report, robots in Japan are now serving as Buddhist priests. The robots are used in temples to perform religious ceremonies, replacing human priests. This marks a shift from human-led to automated ceremonies.
10 source articlesqz.com
WF-XO6C66Happened 12 Jul 2021Live 5 Oct 2026
Userviz machine-learning aimbot shut down after Activision request
A developer known as User101 shut down the Userviz cheat after Activision requested that he stop developing it. The software used computer vision and machine learning to automate aiming in games such as Call of Duty: Warzone, and was marketed as undetectable. It was never published, and the developer said his intention was not to do anything illegal.
Company
User101
AI system
Userviz
7 source articlesvice.com
WF-0XZDESLive 5 Oct 2026
Tencent launches Zero-Point Cruise facial recognition to enforce night-time game curfew
Tencent has introduced a feature called Zero-Point Cruise in its games, which subjects accounts registered as adults that play at night beyond a set time to facial recognition. Anyone who refuses or fails the verification is treated as a minor and logged out. Tencent says this is intended to stop children using adult identities to evade the game curfew, and that adults who mistakenly refuse can wait for the next authentication.
Company
Tencent
AI system
零点巡航 (Zero-Point Cruise)
10 source articlesmp.weixin.qq.com
WF-QSFASRHappened 1 Jan 2019Live 5 Oct 2026
Denmark's Udbetaling Danmark welfare algorithm illegally surveilled millions
Udbetaling Danmark (UDK), Denmark's centralized welfare payment agency, used automated algorithms to check eligibility and detect fraud. The system collected personal data on millions of beneficiaries and their relatives, including non-beneficiaries, in what a think-tank called 'systematic surveillance'. A complaint to the Danish Data Protection Authority led to a ruling that the data collection violated GDPR, and UDK eventually deleted the illegally obtained data. The system also caused erroneous payments, including underpaying 300,000 pensioners by 94 million euros annually.
Company
Udbetaling Danmark
10 source articlesalgorithmwatch.org
WF-UGP6MGHappened 1 Jan 2021Live 5 Oct 2026
Morgan Neville AI use for Anthony Bourdain voice draws backlash
Director Morgan Neville used artificial intelligence to recreate the late Anthony Bourdain's voice for the documentary 'Roadrunner.' Neville stated he intended the AI-generated audio as an Easter egg, but the decision sparked public criticism. He subsequently stated he will never use AI again.
10 source articleswired.com
WF-8LIU7DHappened 15 Mar 2022Live 5 Oct 2026
Federal jury finds Bexar County Sheriff's Flock surveillance-powered traffic stops unconstitutional - San Antonio Current
The system flagged Alek Schott's vehicle as suspicious, leading to a traffic stop, interrogation, and search.
Company
Flock Safety
AI system
Flock
1 source articlesacurrent.com
WF-Y72HNGLive 5 Oct 2026
Bev Standing sues TikTok over unauthorized use of her voice for AI text-to-speech.
Voice actress Bev Standing is suing TikTok, alleging the company used her voice for its text-to-speech feature without compensation or consent. The lawsuit claims that TikTok's AI-generated voice spoke 'foul and offensive language,' causing irreparable harm to her reputation. The case highlights concerns among voice actors about AI voice cloning and the lack of legal protections.
Company
TikTok
AI system
TikTok text-to-speech feature
10 source articlesvice.com
WF-FJUACKLive 5 Oct 2026
Senators question privacy of facial recognition tablets in Uber and Lyft vehicles
U.S. Senators Amy Klobuchar and Richard Blumenthal sent a letter to Uber and Lyft demanding information about interactive tablets installed in vehicles that use facial recognition to target ads. The tablets, provided by company Alfi, use AI to identify passengers' race, gender, and approximate age to serve personalized content. The senators expressed concerns about data collection and privacy, while Alfi claims its automation respects privacy without tracking identifiable information. No specific harm has been reported, but the senators are seeking answers about how the technology is used.
Company
Uber and Lyft
AI system
Alfi interactive tablets with facial recognition AI
10 source articlesklobuchar.senate.gov
WF-SKZU1OHappened 25 May 2021Live 5 Oct 2026
TikTok's beauty filter changed users' faces without their consent
TikTok users reported that the app applied an unwanted beauty filter to their videos, altering the shape of their faces. The effect, which appeared on some Android devices, could not be turned off and caused distress, including gender dysphoria for one user. TikTok acknowledged the issue and said it had been resolved, but did not explain the cause.
Company
TikTok
AI system
TikTok's beauty filter
9 source articlestechnologyreview.com

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

Twitter's image-cropping algorithm found to favor young, thin females

Apple kills plan to detect CSAM in iCloud

Apple Watch algorithm changes raise research data concerns

Social media platforms fail to act on 84% of antisemitic posts

CBSA tested facial recognition on millions at Toronto Pearson Airport in 2016

False AIS tracks simulated for Swedish Navy vessels

Mercadona fined €2.5 million for facial recognition pilot

Virginia crime lab faces first challenge to secret DNA algorithm

Amazon sellers lose thousands after RepricerExpress pricing glitch

UAE cloud seeding drones cause Dubai rainstorms

Doctored audio evidence used to discredit Dubai father in UK custody case

Twitter Japan suspends accounts of government critics

Facebook apologizes after vulgar translation of Chinese leader's name

Pepper robot has severe security vulnerabilities allowing remote takeover

Tapia robot vulnerability at Henn na Hotels allows remote spying on guests.

Robot priests in Japan conduct Buddhist ceremonies

Userviz machine-learning aimbot shut down after Activision request

Tencent launches Zero-Point Cruise facial recognition to enforce night-time game curfew

Denmark's Udbetaling Danmark welfare algorithm illegally surveilled millions

Morgan Neville AI use for Anthony Bourdain voice draws backlash

Federal jury finds Bexar County Sheriff's Flock surveillance-powered traffic stops unconstitutional - San Antonio Current

Bev Standing sues TikTok over unauthorized use of her voice for AI text-to-speech.

Senators question privacy of facial recognition tablets in Uber and Lyft vehicles

TikTok's beauty filter changed users' faces without their consent