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
Une voiture autonome fait un écart soudain et percute... des policiers - Auto Plus
A self-driving car swerved unexpectedly and collided with police officers. The incident was reported by Auto Plus.
石川・小松市の自動運転バスが縁石に接触 3か月前にも同様の事故 市は原因究明進める考え示す（テレビ金沢） - Yahoo!ニュース
An autonomous bus in Komatsu, Ishikawa Prefecture, made contact with a curb. A similar accident occurred three months ago. The city has indicated it will investigate the cause.
Jurnalis Pakistan Alami Kecelakaan Saat Jajal Mobil Otonom - achmadnurhidayat.id
Pakistani journalist Muhammad Aalijah was testing a remote-controlled autonomous car when it lost control and hit a parked police car. The car was operated via satellite and internet, and the connection disruption caused the crash. Aalijah was recording from the passenger seat during the demonstration.
Carro autônomo perde conexão e atinge viatura policial durante teste; assista - Vrum
During a test in Islamabad, Pakistan, an autonomous car lost its internet and satellite connection, causing it to hit a parked police vehicle. The journalist in the passenger seat attempted to engage the handbrake but was unable to avoid the collision. The incident highlighted the risks of testing experimental technology on public roads.
Tumbler Ridge shooting survivors among 30 new plaintiffs suing OpenAI Global News
Amazonが顧客をAIと誤認し、レビューへのアクセス制限 bot被害が拡大 - 日経クロストレンド
Since late 2025, Amazon has restricted some users from seeing more than a few product reviews after misidentifying their visits as AI crawler traffic. Affected customers must email the company and wait several business days for a response. Users are frustrated that Amazon has offered no explanation.
Tesla Meredaksi Data Bantuan Pengemudi dalam Dua Kecelakaan Fatal - readers.id
The website www.readers.id displayed a security verification page instead of article content. The verification process was blocked by incompatible browser extensions or network settings. The page instructs users to disable extensions or check network and firewall settings.
Tesla FSD Masih Tabrak Rel Kereta, Fitur Terbaru Gagal Total - telset.id
Tesla's Full Self-Driving (FSD) v14.3.7 failed to recognize an active railroad crossing and accelerated toward the tracks until the driver manually braked to avoid an oncoming freight train. Dashcam footage of the incident was posted by automotive journalist Lei Xing on August 26, 2026. NHTSA has opened an investigation after receiving 58 reports of FSD traffic violations, including railroad crossing incidents.
Linda, 37, profilerad av Arbetsförmedlingens AI: ”Nedvärderande” - Aftonbladet
Linda Hellman, 37, was profiled by the Swedish Public Employment Service’s AI tool without her knowledge. She felt the experience was degrading and criticized the tool as misleading and unreliable. She has been unemployed for nearly a year and is part of the ‘Rusta och matcha’ program, but has not found a job despite applying for an average of 28 jobs per month.
Kecelakaan Fatal Tesla: Data ADAS Dirahasiakan, Investigasi Terhambat - telset.id
Two fatal Tesla crashes occurred in Arizona and Florida, killing the drivers. In both cases, the Level 2 driver-assist system was active 30 seconds before impact, but Tesla redacted key details from crash reports as confidential business information, hampering the investigation. The NHTSA is now examining how Tesla reports such incidents.
Round Hill demanda a Anthropic y Suno por presunto uso de canciones para entrenar sus modelos de IA - La…
Round Hill Music sued Anthropic and Suno in a California federal court, alleging the AI companies misused hundreds of its copyrighted songs to train their systems. Anthropic allegedly used lyrics from at least 500 songs—including works by James Brown, The Kinks, and the Goo Goo Dolls—to train its Claude chatbot, while Suno used the same songs for its music generation AI. The publisher says it may add thousands more compositions and claims damages could exceed one billion dollars.
AI screening system determined which job applicants to advance to the next stage of recruitment.
OpenAI deactivated the user's account but did not alert police, and the user later committed a mass shooting at Tumbler Ridge Secondary School, killing and wounding students and staff.
Foxglove, supporting the Greater Manchester Coalition of Disabled People, sent a formal pre-action letter to the Department for Work and Pensions over its secret benefit investigations algorithm, demanding the department explain how it works and show evidence it does not discriminate. Campaigners say the algorithm targets disabled people disproportionately and unfairly. Rick Burgess of the GMCDP described how people targeted by the algorithm have sometimes been found dead, starved and dehydrated, having written letters of despair on the same brown envelope the DWP sent them.
Gmail's spam filter marked Republican fundraising emails as spam, preventing them from reaching intended recipients.
Tenant screening companies assign renters a score drawn from data far wider than credit history, an industry subject to less regulation than credit scoring agencies. ProPublica reports that experts warn these algorithms can decide who gets an apartment on the basis of information applicants cannot see or correct. Kim Fuller was denied a rental application on such a score.
The ad delivery system allegedly discriminated in housing ad delivery on the basis of protected characteristics.
A legal action filed in Kenya's High Court claims Meta promoted content that led to ethnic violence and killings in Ethiopia by using an algorithm that prioritises and recommends hateful and violent material on Facebook. The petitioners ask the court to stop the algorithm recommending such content and to compel Meta to create a 200 billion shilling (US$1.6 billion) victims' fund. They are represented by Mercy Mutemi of Nzili and Sumbi Advocates. Amnesty International said Meta must reform its business practices so that Facebook's algorithms do not amplify hatred and fuel ethnic conflict.
Which Facebook users were eligible to receive housing advertisements
Tesla's Autopilot/FSD system controlled the vehicle, leading to multiple fatal crashes.
Navi Mumbai cyber police shut down an alleged fake Apple technical support call centre operating from a rented office at Millennium Business Park in Mahape and arrested four men. The accused posed as Apple customer support executives, telephoned people in the United States and told them their Apple IDs had been compromised. The reporting describes a conventional call-centre fraud and does not identify any AI system as having been used.
The deepfake video impersonated Singapore's Prime Minister to convince a victim to transfer $3.8 million.
AI-generated images were created and circulated online, falsely depicting the hospital visit of a former prime minister.
Displayed deepfake advertisements of Rafał Brzoska to users, leading to financial scams.
Each incident is written from the reporting or the account it cites. A report is an allegation, not a finding.
