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.
Press reportingSelf-reported: a person’s own public post · 10 creditsSelf-reported: told to us directly · 25 credits
Ukraine's Security Service shut down a bot farm in Lviv that used thousands of SIM cards and GSM gateways to manage 18,000 fake accounts. The bots spread false information about bomb threats to cause panic. Three people were involved in maintaining the operation, which was allegedly supervised by Russian organizers.
South Korea's Korea Communications Commission announced policy measures to curb malicious reviews on food delivery apps, following the death of a small restaurant owner in Seoul who allegedly suffered stress after a customer demanded a refund through Coupang Eats. The regulator plans to overhaul review systems, require companies to submit user protection data, and establish a dispute arbitration committee. The article reports that 63.3 percent of restaurant owners in the Seoul area have suffered damage from malicious customer reviews.
ProPublica has reported that companies, including Facebook's own marketing team, used Facebook to show recruitment adverts to specific age groups, alleging age discrimination. Facebook rejects the allegation, saying its adverts are tailored by audience and are part of broader recruitment efforts. The company argues that age-based targeting for employment purposes can be lawful and says it is requiring employers to certify compliance.
MoviePass, the movie subscription service, has announced that its new app will use facial recognition and eye-tracking technology to ensure users are watching adverts. The system, which uses the phone's camera, will pause content if the user looks away. The company says the technology is designed to give advertisers the impact they are looking for. No harm has been reported yet, but the tracking raises privacy concerns.
Ukrainian forces used Turkish-made Bayraktar TB2 drones to strike Russian military convoys and equipment in Ukraine. The drones destroyed multiple vehicles and reportedly caused casualties. The strikes were documented in footage released by the Ukrainian military. The drones' effectiveness against Russian air defences has been noted.
Workers at Sama, a data-labelling company contracted by Facebook to moderate content in Africa, allege they were underpaid (as low as $1.50 per hour), forced to view traumatic material without adequate warning or support, and had their efforts to unionize suppressed. One worker, Daniel Motaung, says he was wrongfully terminated in 2019 after attempting to start a union. A legal NGO, Foxglove, says it is preparing legal action against Sama for wrongful termination. Sama denies any strike or labor action and says it provides competitive salaries and wellness support.
Photographer Jonas Bendiksen created the photobook 'Book of Veles' using 3D software and an OpenAI text generator to depict the North Macedonian town of Veles. The book was published in April 2021 and later screened at the Visa pour l'Image festival without anyone questioning its authenticity. Bendiksen then revealed in an interview with Magnum Photos that the images and text were synthetic. The project raised concerns about the ease of creating convincing misinformation.
Roblox, a children's gaming platform, uses automated and human moderation to review uploaded content. However, sexual content known as "condos" frequently slips through the automated detection system and is only taken down after being discovered. The BBC reported that explicit sexual language also appears in legitimate games. Roblox acknowledged the problem and stated it has zero tolerance for sexual content, but the issue persists.
Twitter said on Wednesday it had mistakenly suspended around a dozen accounts that were posting about Russian military movements. The erroneous action was based on Twitter's rules against synthetic and manipulated media. Twitter has already reinstated access to a number of affected accounts.
Jail and prison officials in at least three states deployed Verus, a software by LEO Technologies, to scan inmate phone calls for keywords related to COVID-19. The system transcribes and analyzes calls, flagging mentions of symptoms or the virus. Advocacy groups argue the surveillance is an abuse of privacy and could lead to retaliation against inmates raising concerns about conditions. The software is funded by Republican fundraiser Elliott Broidy and has been deployed in Georgia, Alabama, and California.
Life360, a family safety app used by 33 million people, sells precise location data of its users to approximately a dozen data brokers, according to interviews with former employees. The company removes obvious identifiers but does not adequately anonymize the data, allowing potential re-identification. Life360 announced it will stop sales to some brokers after the investigation.
Crisis Text Line, a nonprofit mental health support service, uses an AI-driven chat system to collect data from conversations with people in distress. The organization shares anonymized data with its for-profit spinoff, Loris.ai, which uses it to develop customer service software. Critics raise ethical concerns about privacy and consent, though Crisis Text Line asserts the data is stripped of identifying details.
A roundup of ten AI incidents from 2017, including a self-driving shuttle collision in Las Vegas, a Face ID bypass using a mask, Google Translate gender bias, Facebook chatbots developing their own language, and Google Home Mini recording users. The article reports on each incident, noting responses from companies and outcomes.
German authorities used data from a mobile phone contact tracing app, intended for COVID-19 close contact identification, to identify potential witnesses to a crime. Critics say this misuse of the app's data could fuel conspiracy theories about the app and undermine public trust.
Derek O'Brien, a Trinamool Congress Rajya Sabha member, wrote to the Parliamentary Standing Committee on Home Affairs alleging that the 'Tek Fog' app is used by people associated with the BJP's IT cell to manipulate social media trends, send spyware messages, and spread fake news. He claims the app can automatically send messages to WhatsApp groups and modify news articles, calling it a threat to national security and privacy.
The Bulli Bai app posted photos of Muslim women as if for auction, targeting around 100 journalists, activists, and politicians. The app was created by individuals linked to the alt-right group Trads. Police arrested four university students and a web developer. The women reported fear of hate crimes and ongoing harassment.
Amnesty International reports that Hyderabad in Telangana, India, is becoming a total surveillance city with extensive facial recognition technology. The police have been using tablets to stop, search and photograph civilians without charge, and a new Command and Control Centre will connect CCTV cameras for real-time facial recognition. The technology threatens human rights including privacy and freedom of expression, and there is no legislation to protect citizens. Amnesty International is calling for a ban on facial recognition for mass surveillance.
Several local governments in South Korea, including Bucheon, are developing AI systems that use CCTV footage to track individuals for purposes such as disease control, child abuse detection, and personal protection. The systems are being built without obtaining consent from the people captured in the footage, raising concerns about privacy and surveillance. Bucheon plans to launch a smart epidemiological investigation system in January 2022 that will analyze footage from thousands of crime-prevention cameras to track COVID-19 cases and contacts. Critics question the adequacy of de-identification measures and warn of potential misuse of biometric data.
In November 2021, a beta tester reported being sexually harassed by a stranger's avatar in Meta's virtual reality social platform Horizon Worlds. The victim said she was virtually groped and felt isolated when others supported the behavior. Meta acknowledged the incident and said the victim should have used the Safe Zone feature, which they plan to make easier to find. The company stated it is committed to improving safety tools.
Trelleborg municipality in Sweden introduced a fully automated system named Ernst to handle welfare applications. An analysis of the system's source code, obtained via a freedom of information request, revealed that the names and social security numbers of approximately 250 people were included in the code since its introduction, making them visible to anyone who filed a request. The municipality is investigating the leak, which is also the subject of a complaint to the Swedish Parliamentary Ombudsman.
Blind users and disability advocates say AccessiBe, an automated web accessibility overlay, disrupts screen readers on websites, making some sites unnavigable and preventing users from paying rent, shopping or teaching. More than 400 people signed an open letter asking companies to stop using automated overlays. Some blind users have brought ADA lawsuits against companies that use AccessiBe; one case was referred to mediation and another was settled. AccessiBe denies that it was at fault, saying critics do not give specific examples.
A study by KU Leuven and NYU found that Facebook's algorithm misclassifies up to 83% of political ads globally. The system fails to detect undeclared political ads and wrongly labels non-political ads as political, misleading users and advertisers. Facebook has acknowledged the issue but indicated it does not plan to improve detection.
A Denny's restaurant used a Servi robot from Bear Robotics to deliver food to tables, while human servers took orders and served coffee. A viral TikTok video showed the robot, named Janet, delivering hashbrowns and eggs. Customers expressed anger and fear about job loss, with some calling for a boycott. The robot is designed to assist servers, not replace them.
Police surrounded a couple's home and pointed guns at them after a Flock camera lead incorrectly associated their vehicle with a crime.
Each incident is written from the reporting or the account it cites. A report is an allegation, not a finding.
