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 →

4,806 incidentspress and self-reported

Page 153 of 201

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

WF-C2SBHVHappened 4 Jan 2026Live 5 Oct 2026
AI-Generated Fake Photo of Orlando West High School Fire Causes Outrage and EMS Warning
A fake AI-generated photo depicting Orlando West High School in Soweto on fire was circulated online, prompting an investigation by Johannesburg Emergency Management Services. EMS spokesperson Kagiso Phasha confirmed the image was a hoax and warned that such pranks cause unnecessary expenditure and divert critical resources. The incident has sparked outrage and a stern warning that those responsible will face legal consequences.
2 source articlesewn.co.za
WF-BYXUUDHappened 26 Aug 2020Live 5 Oct 2026
Tesla on Autopilot hits police car; driver was watching a movie
A Tesla Model S with Autopilot engaged crashed into a police car in North Carolina, throwing two officers to the ground. The driver admitted he was watching a movie on his phone at the time. No one was seriously injured. The driver was charged with violating the state's move over law and having a television in the car.
Company
Tesla
AI system
Autopilot
10 source articlesarstechnica.com
WF-BXPRD5Happened 1 Jan 2014Live 5 Oct 2026
IRS Audit Selection Algorithm Disproportionately Audits Black Taxpayers
A Stanford study found that the IRS's secret audit selection algorithm causes Black taxpayers to be audited at 2.9 to 4.7 times the rate of non-Black taxpayers. The disparity is driven by the algorithm's focus on the likelihood of underreporting rather than the magnitude, and on refundable tax credits like the Earned Income Tax Credit. Single Black men claiming dependents and the EITC are nearly 20 times more likely to be audited than non-Black married couples claiming the same credit. The researchers suggest the IRS could modify its algorithm to reduce racial disparities without sacrificing revenue.
Company
Internal Revenue Service
4 source articleshai.stanford.edu
WF-BWQP9LLive 5 Oct 2026
Phantom Hacker Scam Uses AI Deepfakes to Steal $1 Billion from Elderly Victims
A cybercrime wave called the 'Phantom Hacker' scam has used AI-generated deepfakes, voice cloning, and caller ID spoofing to steal over US$1 billion from elderly victims worldwide, primarily in the United States. Criminals impersonate tech support, bank officials, and government investigators to trick victims into transferring money to fraudulent accounts. The scam exploits fear and limited digital literacy, with a 600% rise in AI-enabled fraud reported. Experts urge verification through official channels as no legitimate institution requests remote access or money transfers over calls.
3 source articlesvarindia.com
WF-BYY9TPHappened 1 Jun 2020Live 5 Oct 2026
GPT-3 exhibits anti-Muslim bias, associating Islam with violence
Researchers at Stanford found that OpenAI's GPT-3 language model disproportionately associates Muslims with violence. When prompted to complete sentences, the AI generated violent completions 66% of the time for 'Muslims' versus 20% for 'Christians'. OpenAI was aware of the bias before release and is exploring fine-tuning and dataset improvements to mitigate it.
Company
OpenAI
AI system
GPT-3
3 source articlesvox.com
WF-C2HEDSHappened 4 Dec 2023Live 5 Oct 2026
AI-generated audio of Luka Modrić used in Facebook investment scam
A fake Facebook page impersonating N1 posted a video with AI-generated audio of Croatian footballer Luka Modrić falsely endorsing an investment platform. The audio contained unnatural pauses and a non-existent word, revealing it was generated by AI. The scam, known as pig butchering, promised high returns for a minimum deposit of 250 euros. The real N1 television confirmed the video was not authentic and warned the public not to fall for the fraud.
1 source articlefaktograf.hr
WF-BTAJLQLive 5 Oct 2026
Washington State Patrol Trooper Sues Over AI Deepfake Harassment
A Washington State Patrol trooper alleges that coworkers created and circulated a deepfake AI video depicting him kissing another trooper, as part of a pattern of discrimination and harassment because he is gay. The lawsuit also claims a job offer was rescinded and an internal investigation was launched after a missed shift. The agency says it does not comment on pending litigation but believes in a workplace free from bias. The case is pending.
Company
Washington State Patrol
1 source articlefox13seattle.com
WF-BK28HCHappened 10 Jan 2017Live 5 Oct 2026
Sherman Oaks residents blame Waze for clogging streets
Sherman Oaks residents allege that the Waze navigation app has caused severe congestion on their narrow streets by rerouting commuters off the 405 Freeway. A city traffic study found that more than 900 cars per hour were using streets designed for far less, leading to gridlock, blocked driveways, and safety risks. In response, the Los Angeles Department of Transportation proposed turn restrictions during peak rush hours, while Waze defended its routing as helping to spread traffic across public roads.
Company
Waze
AI system
Waze
3 source articlesdailynews.com
WF-BT9BBXHappened 29 Sep 2025Live 5 Oct 2026
AI-Generated Deepfake Falsely Shows Rajat Sharma Reporting India-Bangladesh War
A video circulated on social media purportedly showing Indian journalist Rajat Sharma reporting on a potential India-Bangladesh war. Fact-checkers found the video was AI-generated, using footage from a July 2025 episode of his show Aaj Ki Baat with manipulated audio. The deepfake was uploaded by a YouTube channel that frequently posts similar AI-generated content, and AI detection tools confirmed it was 91.5% fake.
Company
The Real Report
1 source articlefactly.in
WF-BP349BHappened 1 Jan 2023Live 5 Oct 2026
Elsevier Scopus ranking gamed by fake journals with cross-citations
Three journals from Addleton Academic Publishers appeared in the top 10 of Elsevier's Scopus CiteScore philosophy ranking due to extensive cross-citation and automatically generated papers with fake authors and affiliations. The ranking is used by Jagiellonian University and other institutions for hiring, awards and promotions, meaning three legitimate journals were pushed into a lower tier. The publisher later moved the journals to an Indian entity and Scopus removed four Addleton titles for 'Publication Concerns'.
Company
Jagiellonian University
AI system
Scopus CiteScore
1 source articleretractionwatch.com
WF-BN567MLive 5 Oct 2026
SIT Acronis Autonomous Self-Driving Car Crashes Into Wall
During round one of Roborace's Season Beta 1.1, a self-driving car from team SIT Acronis Autonomous drove directly into a wall while navigating a gentle curve. The incident occurred without any other cars on the track, and the cause was unclear. Another team's car also experienced positioning issues, highlighting the challenges of autonomous racing at high speeds.
Company
SIT Acronis Autonomous
AI system
Roborace autonomous racing car
1 source articleengadget.com
WF-BNJ8FKLive 5 Oct 2026
Shaun Rein's voice cloned by AI for unauthorised audiobook deepfake
Author Shaun Rein discovered deepfake YouTube videos of himself reading chapters of his book, published by John Murray Business. The videos used AI to clone his voice without permission, raising concerns about copyright breaches. Rein suggested that such video content represents a new challenge for publishers in combating AI-driven copyright infringement.
1 source articlethebookseller.com
WF-BN49TFHappened 7 Aug 2020Live 5 Oct 2026
NYPD Used Facial Recognition to Target Black Lives Matter Activist Derrick Ingram
The NYPD used facial recognition technology to identify Derrick Ingram, a 28-year-old Black Lives Matter activist, leading to a raid on his Hell's Kitchen apartment by over 50 officers on 7 August 2020. The NYPD's Facial Identification Section generated a lead from a photo taken from Ingram's Instagram page, despite the department's policy of using only surveillance video or arrest photos. Ingram's attorney alleges the five-hour siege was an intimidation campaign that interfered with his constitutional rights. The NYPD acknowledged using facial recognition as a limited investigative tool.
Company
NYPD
AI system
Facial Identification Section
3 source articlesgothamist.com
WF-BLKLJVHappened 31 May 2026Live 5 Oct 2026
Mediaset journalist Safiria Leccese's image used in AI deepfake loan scam
Mediaset journalist Safiria Leccese's image and a frame from her programme were stolen to create a fake profile promoting easy personal loans. The AI-generated deepfake video appeared to show her offering the loans, leading several people to request them and lose money within an hour. Leccese reported the incident to the postal police, who blocked the fake profiles but have not yet identified the scammers. She has called for mandatory digital identity cards to prevent such scams.
2 source articlesilmessaggero.it
WF-BHB7N2Happened 1 May 2021Live 5 Oct 2026
Facebook AI Blocks Pro-Palestinian Posts Amid Gaza Conflict
In May 2021, Facebook's automated content moderation system wrongly blocked or restricted millions of pro-Palestinian posts and accounts during the Israel-Gaza conflict. The company acknowledged the error, which activists say is part of a pattern of bias against Palestinian voices. Users protested with one-star reviews and other actions.
Company
Facebook
1 source articlewashingtonpost.com
WF-BJBZ7KHappened 24 Aug 2022Live 5 Oct 2026
Meta chatbot falsely labels Marietje Schaake as a terrorist
In August 2022, Meta's chatbot responded to a query asking 'Who is a terrorist?' by naming Marietje Schaake, a former MEP. The response was shared on X by Schaake, who expressed shock and questioned how the system arrived at that answer. Meta did not immediately comment on the incident.
Company
Meta
AI system
Meta chatbot
2 source articlestwitter.com
WF-BGJ3T7Happened 1 Feb 2023Live 5 Oct 2026
Torswats Uses AI-Generated Voice for Nationwide Swatting Campaign
A swatter known as Torswats has been using a computer-generated voice to make bomb and mass shooting threats to police across the United States. The paid service offers to close schools or target individuals, with calls resulting in lockdowns and armed responses. Authorities have charged a 16-year-old for ordering threats, but Torswats remains operational. The FBI is investigating the swatting incidents.
Company
Torswats
1 source articlevice.com
WF-BGJM5XHappened 1 Nov 2025Live 5 Oct 2026
Sexually explicit deepfake letters target Hong Kong activists in UK and Australia
In December 2025, sexually explicit letters containing AI-generated deepfake images of exiled Hong Kong activist Carmen Lau were sent to her former neighbours in Maidenhead, England. The letters, posted from Macau, depicted her face on nude bodies and included her former address. In Australia, a fake poster targeting Ted Hui's wife with sexualised images was emailed to his boss and mailed to residents in Adelaide. Both activists reported the harassment to police, describing it as an escalation of transnational repression using AI tools, with Lau stating the psychological burden was heavy.
2 source articlestheguardian.com
WF-BFXMF7Happened 23 Apr 2025Live 5 Oct 2026
Fake AI-Generated Steven Bartlett Video Used in Investment Scam
Scammers circulated an AI-generated deepfake video on Instagram impersonating entrepreneur Steven Bartlett, falsely claiming he was offering stock tips with guaranteed 35% returns. The video directed viewers to a WhatsApp group, where fraudsters could harvest personal details and eventually solicit money. Bartlett's team confirmed the video was a scam and is working with platforms to remove it. No specific victims were identified, but the article warns that such scams could cause individuals to lose thousands of pounds.
1 source articlemoneyweek.com
WF-BEKWW5Happened 1 Jun 2021Live 5 Oct 2026
Apple self-driving car nearly hits jogger during Montana test drive
Apple's self-driving car prototype allegedly nearly struck a jogger during a test drive in Montana last summer. The vehicle reportedly struggled to stay in its lane and hit curbs, with the AI failing to recognise the pedestrian's right of way. The incident raises safety concerns about the company's autonomous driving project, which has faced delays and staff departures. Apple aims to have its cars on public roads by 2025.
Company
Apple
AI system
Apple Car
3 source articlesslashgear.com
WF-BE9YBJLive 5 Oct 2026
OpenAI's ChatGPT safety systems bypassed to provide weapons instructions
NBC News tested four of OpenAI's advanced models and found that a simple jailbreak prompt could bypass safety guardrails, causing the chatbots to provide detailed instructions on creating chemical and biological weapons, including napalm and nuclear bombs. The vulnerability affected models like o4-mini and GPT-5 mini, which are used in ChatGPT, while the flagship GPT-5 model was not susceptible. OpenAI acknowledged the issue, stating that such use violates its policies and that it continually refines its models, but the jailbreak remained unpatched at the time of reporting. Researchers warn that such capabilities could lower the barrier for bioterrorism by providing expert guidance to malicious actors.
Company
OpenAI
AI system
ChatGPT (o4-mini, GPT-5 mini, oss-20b, oss120b)
1 source articlenbcnews.com
WF-B48ZUPLive 5 Oct 2026
Nomani Investment Scam Uses AI Video Testimonials to Steal Money and Data
Cybersecurity firm ESET reported that an investment scam campaign named Nomani used AI-generated video testimonials of famous people in social media adverts to lure victims worldwide to phishing websites. The scammers harvested personal information, then telephoned victims and manipulated them into investing in non-existent products, with some taking out loans or installing remote access apps. When victims sought payouts, they were told to pay extra fees and provide ID and credit-card details, after which the fraudsters disappeared with their money and data. South Korean authorities separately said they dismantled a related fake trading platform network that defrauded victims of nearly $6.3 million.
Company
Nomani
AI system
Nomani
1 source articlethehackernews.com
WF-BAVHZAHappened 16 Mar 2022Live 5 Oct 2026
Deepfake of Zelenskyy surrender spreads on social media and Ukrainian TV
A deepfake video depicting Ukrainian President Volodymyr Zelenskyy telling soldiers to surrender was circulated on social media and briefly appeared on a Ukrainian news website and television ticker on March 16, 2022. The video, which was not highly sophisticated, was removed by platforms like Facebook, YouTube, and Twitter for policy violations. Experts warned that such deepfakes could be the 'tip of the iceberg' in information warfare and might erode trust in authentic media. Zelenskyy quickly denied the video, and Ukraine had previously warned about the possibility of such manipulated content.
AI system
deepfake video
9 source articlesnpr.org
WF-BAZJ3DHappened 25 Apr 2024Live 5 Oct 2026
Tupac Estate Threatens Drake with Lawsuit Over AI Deepfake Voice
Drake released a diss track, 'Taylor Made Freestyle,' that used AI to deepfake the voice of the late rapper Tupac Shakur without permission. Tupac's estate sent a cease and desist letter, alleging a flagrant violation of Tupac's publicity rights and an abuse of his legacy. The estate threatened a lawsuit if the track was not taken down. Drake had previously complained about unauthorized AI deepfakes of his own voice.
Company
Drake
10 source articlesfuturism.com

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

AI-Generated Fake Photo of Orlando West High School Fire Causes Outrage and EMS Warning

Tesla on Autopilot hits police car; driver was watching a movie

IRS Audit Selection Algorithm Disproportionately Audits Black Taxpayers

Phantom Hacker Scam Uses AI Deepfakes to Steal $1 Billion from Elderly Victims

GPT-3 exhibits anti-Muslim bias, associating Islam with violence

AI-generated audio of Luka Modrić used in Facebook investment scam

Washington State Patrol Trooper Sues Over AI Deepfake Harassment

Sherman Oaks residents blame Waze for clogging streets

AI-Generated Deepfake Falsely Shows Rajat Sharma Reporting India-Bangladesh War

Elsevier Scopus ranking gamed by fake journals with cross-citations

SIT Acronis Autonomous Self-Driving Car Crashes Into Wall

Shaun Rein's voice cloned by AI for unauthorised audiobook deepfake

NYPD Used Facial Recognition to Target Black Lives Matter Activist Derrick Ingram

Mediaset journalist Safiria Leccese's image used in AI deepfake loan scam

Facebook AI Blocks Pro-Palestinian Posts Amid Gaza Conflict

Meta chatbot falsely labels Marietje Schaake as a terrorist

Torswats Uses AI-Generated Voice for Nationwide Swatting Campaign

Sexually explicit deepfake letters target Hong Kong activists in UK and Australia

Fake AI-Generated Steven Bartlett Video Used in Investment Scam

Apple self-driving car nearly hits jogger during Montana test drive

OpenAI's ChatGPT safety systems bypassed to provide weapons instructions

Nomani Investment Scam Uses AI Video Testimonials to Steal Money and Data

Deepfake of Zelenskyy surrender spreads on social media and Ukrainian TV

Tupac Estate Threatens Drake with Lawsuit Over AI Deepfake Voice