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 →

4,498 incidentspress and self-reported

Page 106 of 188

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

WF-UFX4OLLive 5 Oct 2026
Researchers demonstrate drone attack on Renault Captur ADAS with fake road signs
Security researchers from Ben Gurion University demonstrated a vulnerability in the Mobileye ADAS system used in a Renault Captur. They used drones to project fake road signs for 100 milliseconds, too quick for human perception but detectable by the system. The attack could trick the car into making unsafe maneuvers, though the system is only Level 0 (advisory). The researchers published their findings in a paper titled 'MobilBye: Attacking ADAS with Camera Spoofing'.
AI system
Mobileye ADAS
7 source articlesboingboing.net
WF-HH2ZL6Live 5 Oct 2026
Epic sepsis prediction algorithm misses two-thirds of cases, study finds
A study of nearly 30,000 patients at University of Michigan hospitals found that Epic Systems' sepsis prediction algorithm missed two-thirds of sepsis cases and frequently issued false alarms. The algorithm, used by over 100 US health systems, is designed to alert staff to potential sepsis, but the study suggests it provides little value and may contribute to alert fatigue. Epic disputed the study's conclusions, claiming its system has helped save lives.
Company
Epic Systems
AI system
Epic sepsis prediction model
6 source articleswired.com
WF-OH0YGMHappened 1 Oct 2020Live 5 Oct 2026
Microsoft removes user names from Productivity Score after privacy backlash
Microsoft rolled out its Productivity Score feature in Microsoft 365 in October 2020, which allowed managers to see individual employee data on email, chat, and meeting usage. Privacy experts criticized it as a workplace surveillance tool. After backlash, Microsoft removed user names and changed the feature to aggregate data at the organization level only.
Company
Microsoft
AI system
Productivity Score
8 source articlesgeekwire.com
WF-JWRXGYHappened 7 Jul 2020Live 5 Oct 2026
Cense exposed 2.5 million records of auto accident victims online
On July 7, 2020, a security researcher discovered 2.5 million records containing personal and medical data of auto accident victims exposed online. The records, belonging to New York-based AI company Cense, included names, insurance policy numbers, claim numbers, and medical diagnosis notes. The data was labeled as staging data, possibly intended for temporary storage before being loaded into an AI system. After the researcher sent a responsible disclosure notice, Cense restricted public access to the database.
Company
Cense
AI system
Cense
5 source articlessecuritydiscovery.com
WF-DCT085Happened 1 Sep 2019Live 5 Oct 2026
Swedish police fined for unlawful use of Clearview AI facial recognition
Sweden's data protection authority IMY fined the Swedish Police Authority €250,000 for unlawfully using Clearview AI's facial-recognition app. The police used the app between autumn 2019 and March 2020 without authorisation or a required data protection impact assessment. The IMY ordered the police to ensure Clearview AI deletes the data and to train employees to avoid future breaches.
Company
Swedish Police Authority
AI system
Clearview AI
3 source articlescomputerweekly.com
WF-GTGQFRLive 5 Oct 2026
Banjo's Live Time system overstated capabilities and posed privacy risks for Utah public safety data
The Utah State Auditor reviewed the contract for Banjo's Live Time public safety application. The system was found to lack advertised AI technology and posed security risks due to direct database queries to PSAP databases. Banjo also overstated its capabilities in its response to the RFP. The auditor concluded that personally identifiable information was unlikely accessed but recommended more rigorous vetting and better security practices.
Company
Utah Attorney General's Office
AI system
Live Time
5 source articless3.documentcloud.org
WF-QSQO1HLive 5 Oct 2026
Trivago loses Australian appeal over misleading hotel rate rankings
The Federal Court of Australia has dismissed Trivago's appeal against a ruling that the online travel company breached Australian consumer law. The court found that Trivago's website used an algorithm that gave prominence to hotels paying higher fees, so the most prominent offers were not always the cheapest for consumers. Consumers may therefore have paid more for rooms, and hotels lost direct bookings. The case returns to the Federal Court for consideration of penalties and other orders sought by the Australian Competition and Consumer Commission.
Company
Trivago
10 source articlestravelweekly.com.au
WF-I2B6P1Happened 31 Oct 2019Live 5 Oct 2026
Guo Bing sues Hangzhou safari park over mandatory face-scans
Guo Bing, a legal academic, filed a lawsuit against a local safari park in Hangzhou after the park informed season-pass holders that admission would require a face-scan. He claimed the requirement violated his privacy. The lawsuit is the first of its kind in China and has generated significant public attention.
10 source articleseconomist.com
WF-U19DETHappened 1 Oct 2015Live 5 Oct 2026
Uber sued over alleged racially biased star rating system
A former Uber driver in San Diego filed a federal lawsuit alleging that Uber's star rating system, based on passenger reviews, discriminates against drivers who are not white or who have accents. The suit claims Uber terminates drivers whose average ratings fall too low, and that this disproportionately affects minority drivers. Uber denies the allegation, stating that ridesharing has reduced bias. The plaintiff, Thomas Liu, an Asian driver from Hawaii, was fired in October 2015 after his rating dropped below 4.6.
Company
Uber
AI system
star rating system
10 source articlesnpr.org
WF-FA32HOHappened 1 Mar 2019Live 5 Oct 2026
Tesla Model 3 crash with truck kills driver using partial automation
At 6:17 a.m. on March 1, 2019, a Tesla Model 3 operating with partial driving automation struck a semitrailer in Delray Beach, Florida, killing the 50-year-old driver. The NTSB determined the probable cause was the truck driver's failure to yield right of way combined with the car driver's inattention due to overreliance on the automation, with Tesla's system design allowing disengagement as a contributing factor.
Company
Tesla
10 source articlesntsb.gov
WF-ITDLNTHappened 1 Nov 2018Live 5 Oct 2026
Uber and Lyft pricing algorithms charge more in non-white areas
A study of over 100 million rides in Chicago between November 2018 and December 2019 found that Uber and Lyft's pricing algorithms charged higher fares per mile for trips to or from neighbourhoods with higher proportions of non-white residents. The researchers analysed trip data against census demographics and found no statistical link to higher demand in those areas. Both companies acknowledged the study and said they would review the findings, but no changes have been reported.
Company
Uber
AI system
Uber pricing algorithm
10 source articlesnewscientist.com
WF-DSDQLLHappened 9 Aug 2019Live 5 Oct 2026
Tesla Model 3 crashes into tow truck on Autopilot in Moscow
A Tesla Model 3 crashed into a stationary tow truck on a highway near Moscow while reportedly on Autopilot. The driver, Alexey Tretyakov, and his two children were injured and taken to hospital. The vehicle caught fire and exploded. Police believe the Autopilot system failed to recognize the tow truck. The driver confirmed he was using Autopilot at the time.
Company
Tesla
AI system
Autopilot
9 source articleselectrek.co
WF-UOKPM7Live 5 Oct 2026
Hackers trick Tesla Autopilot into veering into oncoming lane
Researchers at Tencent's Keen Security Lab demonstrated that stickers placed on a road could trick a Tesla Model S's Autopilot system into veering into the oncoming traffic lane. The attack exploited the system's reliance on camera-based computer vision to recognise lanes. Tesla responded that this vulnerability was unrealistic because a driver can easily override Autopilot, while saying a separate steering-wheel hijack had been fixed in a software update.
Company
Tesla
AI system
Autopilot
9 source articlestechnologyreview.com
WF-FRRPXYHappened 1 Dec 2019Live 5 Oct 2026
Kneron researchers bypass facial recognition at airport and train station using mask and photo
In December 2019, Kneron researchers tested facial recognition systems at a high-speed rail station in China, Schiphol Airport, point-of-sale terminals, and a mobile phone. Using a high-quality 3D mask and a photo on a phone screen, they successfully deceived the systems, gaining access and making purchases. The survey highlighted security shortfalls in many commercial facial recognition solutions.
Company
Schiphol Airport; high-speed rail station in China
10 source articleskneron.com
WF-HOQGL3Happened 1 May 2016Live 5 Oct 2026
King's Cross developer scraps facial recognition after public outcry
The developer of the King's Cross site in London admitted it had used facial recognition software in two cameras on a busy pedestrian street between May 2016 and March 2018 without public consent. Following a backlash and questions from the Information Commissioner's Office, the developer announced it would not deploy the technology in the future and had abandoned plans for wider use. The system was used to help police detect crime, but critics highlighted the lack of transparency and privacy concerns.
Company
King's Cross development (Argent-led consortium)
10 source articlestheguardian.com
WF-SASGFEHappened 1 Sep 2018Live 5 Oct 2026
Anderstorp gymnasium trials facial recognition to register student attendance
Anderstorp gymnasium in Skellefteå, Sweden, used a facial recognition system for a few weeks in autumn 2018 to automatically register students' attendance when they entered a classroom. The system, made by an unnamed manufacturer, replaced teachers' manual attendance records for one class. Pupils were mostly satisfied, and the school's principal described the technology as safe. Skellefteå municipality and the manufacturer are discussing a possible continuation, with the manufacturer recommending consultation with Datainspektionen before scaling up.
Company
Skellefteå kommun
10 source articlessvt.se
WF-CEA1N6Happened 1 Jan 2018Live 5 Oct 2026
AI voice bots used for telemarketing in China raise privacy concerns
In China, companies use AI-powered voice bots to cold-call consumers, often without disclosing that the caller is a bot. The bots can respond flexibly and sometimes know the recipient's name, leading to privacy concerns. The practice has generated numerous complaints to the Ministry of Industry and Information Technology, and lawyers say it violates China's privacy laws.
8 source articlessixthtone.com
WF-02DLSMHappened 24 Jun 2019Live 5 Oct 2026
Sidewalk Labs releases Toronto waterfront plan amid privacy concerns
Sidewalk Labs, a subsidiary of Alphabet, released a 1,500-page plan for a new development on Toronto's eastern waterfront, including the Quayside site. The plan involves extensive data collection and management, raising privacy concerns from critics and the public. Waterfront Toronto, the overseeing body, is reviewing the plan and has expressed concerns about its scope and data governance. The company has proposed an independent trust to oversee data, but critics remain unsatisfied.
Company
Sidewalk Labs
AI system
Quayside development
10 source articlescbc.ca
WF-G4XV5BHappened 23 Jul 2019Live 5 Oct 2026
MIT-IBM Watson AI Lab's AI Portrait Ars produces whitewashed portraits of people of colour
AI Portrait Ars, developed by researchers at the MIT-IBM Watson AI Lab, is reported to have generated Renaissance-style portraits that lightened the skin and altered the facial features of people of colour. The tool, trained on tens of thousands of paintings from the Western artistic tradition, was criticised for reproducing that bias in its data set. The creators acknowledged the bias but did not respond to a request for comment.
Company
MIT-IBM Watson AI Lab
AI system
AI Portrait Ars
10 source articlesmashable.com
WF-961A0LHappened 1 Dec 2018Live 5 Oct 2026
Educational Testing Service's E-rater algorithm biases essay scores against minority students
The Educational Testing Service's E-rater algorithm, used to grade essays on the GRE and other standardized tests, has been found to systematically give higher scores to students from mainland China and lower scores to African American students compared to human graders. The bias stems from the algorithm's reliance on surface-level metrics like vocabulary and sentence length, which disadvantage certain groups. Despite studies dating back to 1999, the bias persists, and in many states, only a small percentage of essays are reviewed by humans.
Company
Educational Testing Service
AI system
E-rater
10 source articlesvice.com
WF-SJJF4RHappened 22 Aug 2019Live 5 Oct 2026
Deepfake voice generator impersonates Jordan Peterson without consent
A website called notjordanpeterson.com released an AI engine that allows users to generate audio in Jordan Peterson's voice from any text input. The system was created using a sample of Peterson's voice from public lectures. Peterson expressed concern about potential misuse for fraud and defamation, and the site subsequently disabled its functionality in response to his objections.
Company
notjordanpeterson.com
10 source articlesjordanbpeterson.com
WF-1O81PGHappened 1 Jan 2017Live 5 Oct 2026
Newham Council fined £145,000 over leaked gang matrix data
Newham Council was fined £145,000 by the Information Commissioner's Office after a leaked unredacted gangs matrix, containing details of 203 suspected gang members, ended up in the hands of rival gang members via Snapchat. The leak occurred in January 2017 when a council employee emailed both redacted and unredacted versions to 44 recipients. The ICO found the breach was unnecessary and that the council failed to report it promptly. The council apologised and accepted the breach was not deliberate.
Company
Newham Council
AI system
Gangs matrix
8 source articlesnewhamrecorder.co.uk
WF-C9VJNAHappened 1 Dec 2018Live 5 Oct 2026
CNIL says facial recognition trial in French high schools is illegal
The French data protection authority (CNIL) issued a non-binding opinion that a planned trial of facial recognition at two high schools in Nice and Marseille is illegal under the GDPR. The trial, approved by the South Region in December 2018, involved biometric portals to identify students. Digital rights groups, parents and teachers' unions opposed the experiment and filed a lawsuit. The CNIL stated that the system was too intrusive and that less intrusive methods could achieve the same purpose.
Company
South Region (Région Sud)
AI system
Cisco facial recognition system
10 source articlespolitico.eu
WF-1HWBOZHappened 26 May 2019Live 5 Oct 2026
MIT researchers used YouTube videos to train Speech2Face without consent
MIT researchers used YouTube videos, including one of Nick Sullivan, to train a model called Speech2Face that generates faces from voices. Sullivan discovered his data was used without his consent and expressed surprise. The researchers defended their use of public data, but the incident raised ethical questions about consent in AI research.
Company
Massachusetts Institute of Technology
AI system
Speech2Face
10 source articlesslate.com

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

Researchers demonstrate drone attack on Renault Captur ADAS with fake road signs

Epic sepsis prediction algorithm misses two-thirds of cases, study finds

Microsoft removes user names from Productivity Score after privacy backlash

Cense exposed 2.5 million records of auto accident victims online

Swedish police fined for unlawful use of Clearview AI facial recognition

Banjo's Live Time system overstated capabilities and posed privacy risks for Utah public safety data

Trivago loses Australian appeal over misleading hotel rate rankings

Guo Bing sues Hangzhou safari park over mandatory face-scans

Uber sued over alleged racially biased star rating system

Tesla Model 3 crash with truck kills driver using partial automation

Uber and Lyft pricing algorithms charge more in non-white areas

Tesla Model 3 crashes into tow truck on Autopilot in Moscow

Hackers trick Tesla Autopilot into veering into oncoming lane

Kneron researchers bypass facial recognition at airport and train station using mask and photo

King's Cross developer scraps facial recognition after public outcry

Anderstorp gymnasium trials facial recognition to register student attendance

AI voice bots used for telemarketing in China raise privacy concerns

Sidewalk Labs releases Toronto waterfront plan amid privacy concerns

MIT-IBM Watson AI Lab's AI Portrait Ars produces whitewashed portraits of people of colour

Educational Testing Service's E-rater algorithm biases essay scores against minority students

Deepfake voice generator impersonates Jordan Peterson without consent

Newham Council fined £145,000 over leaked gang matrix data

CNIL says facial recognition trial in French high schools is illegal

MIT researchers used YouTube videos to train Speech2Face without consent