AI Incident Database
Browse AI incidents
Every incident record (metadata only) from the weekly snapshot of 2026-09-07. Filter, then export the selection with its licence attached; each row links to the full record and its reports.
Filter incidents
-
Fake AI-Generated Law Firms Sent Fake DMCA Notices to Increase SEO AIID ↗
In March 2024, fake law firms using AI-generated identities sent fraudulent DMCA takedown notices to website owners, demanding backlinks for SEO gains. These AI-generated law firms, like "Commonwealth Legal," used GAN models for realistic attorney images and fabricated bios. The scam involved fake legal threats to coerce site owners into adding backlinks, exploiting AI technology for deceptive practices.
-
Over 400 Purportedly AI-Driven Scams Reportedly Led to $8M Loss for Australians in 2023 AIID ↗
In 2023, Australians reportedly lost over $8 million to scams involving purported deepfake videos and fake news articles that falsely endorsed investment trading platforms. Scammers reportedly used AI-generated content featuring celebrities to mislead victims, leading to significant financial losses. The National Anti-Scam Centre received over 400 reports of these incidents. One man is reported to have lost over $80,000 in cryptocurrency.
-
Russia-Linked AI CopyCop Site Identified as Modifying and Producing at Least 19,000 Deceptive Reports AIID ↗
In early March 2024, a network named CopyCop began publishing modified news stories using AI, altering content to spread partisan biases and disinformation. These articles, initially from legitimate sources, were manipulated by AI models, possibly developed by OpenAI, to disseminate Russian propaganda. Over 19,000 articles were published, targeting divisive political issues and creating false narratives.
-
Character.ai Chatbot Allegedly Influenced Teen User Toward Suicide Amid Claims of Missing Guardrails AIID ↗
A 14-year-old, Sewell Setzer III, died by suicide after reportedly becoming dependent on Character.ai's chatbot, which engaged him in suggestive and seemingly romantic conversations, allegedly worsening his mental health. The chatbot, personified as a fictional Game of Thrones character, reportedly encouraged harmful behaviors, fueling his obsessive attachment. The lawsuit claims Character.ai lacked safeguards to prevent vulnerable users from forming dangerous dependencies on the AI.
-
Snapchat's Algorithm Alleged to Link Minor with Sex Offenders AIID ↗
A judge ruled Snapchat not liable under Section 230 after its algorithm connected a minor with convicted sex offenders on multiple occasions, leading to sexual assaults first in 2019 and again in 2021. The platform's "Quick Add" feature was implicated in facilitating the connections between the minor and the offenders.
-
Seeming Pattern of Gemini Bias and Sociotechnical Training Failures Harm Google's Reputation AIID ↗
Google's Gemini chatbot faced many reported bias issues upon release, leading to a variety of problematic outputs like racial inaccuracies and political biases, including regarding Chinese and Indian politics. It also reportedly over-corrected racial diversity in historical contexts and advanced controversial perspectives, prompting a temporary halt and an apology from Google.
-
ChatGPT Glitch Disrupts User Interactions with Nonsensical Outputs AIID ↗
ChatGPT experienced a bug causing it to produce unexpected and nonsensical responses, leading to widespread reports of user confusion and concern. OpenAI identified and fixed the language processing bug, restoring normal service.
-
AI incident: sexual content (Feb 2024) AIID ↗
An AI Incident Database record involving sexual content, dated 20 February 2024. Risk classification, related incidents and the laws on record for this risk area.
-
Purported AI-Cloned Voice Depicts Kenyan President William Ruto Speaking French in Viral Video AIID ↗
A purportedly AI-manipulated video widely shared on TikTok and Facebook presents Kenyan President William Ruto speaking in French during a speech in Tokyo on February 7, 2024. A reported reverse image search confirmed the original footage shows Ruto speaking in English to the Kenyan diaspora. The reportedly manipulated version used AI-generated French audio over the original video, making it a case of voice cloning rather than a full audiovisual deepfake.
-
Deepfakes Circulating and Eroding Electoral Integrity in the Lead-up to 2024 South Korean legislative election AIID ↗
In the lead-up to Korea's parliamentary elections, at least 129 deepfake videos and images were reported to have been detected, violating new election laws. These AI-generated deepfakes were used to mislead and manipulate public opinion, prompting a crackdown by the National Election Commission.
-
Alleged State-Sponsored Hackers Escalate Purported Phishing Attacks Using Artificial Intelligence AIID ↗
State-sponsored hackers from North Korea, Iran, Russia, and China are reportedly leveraging artificial intelligence to conduct sophisticated phishing and social engineering attacks. They target global defense, cybersecurity, and cryptocurrency sectors, aiming to steal sensitive information and, in the case of North Korea, cryptocurrencies to help fund its illicit nuclear program.
-
Deepfake Generated by the Lincoln Project of Trump's Father Used in Political Attack Ad AIID ↗
The Lincoln Project used AI to create a deepfake video of Donald Trump's deceased father criticizing him. Although they made it clear that the video was a deepfake, the deeply personal nature of the attack represents a corrosive use of artificial intelligence in undermining democratic norms during an election cycle.
-
York County, South Carolina, Deputies Reportedly Held Local Motorist at Gunpoint After Flock Safety License-Plate Reader Misread AIID ↗
Steven Melvin, a 42-year-old York County, South Carolina, motorist was reportedly detained at gunpoint for more than five minutes after a Flock Safety automated license-plate reader misidentified the first letter of his plate and matched his BMW to a similar stolen vehicle. The sheriff’s office later said an illegal tinted plate cover caused the malfunction and found the deputy’s actions “reasonable”; Flock said alerts should be independently verified before enforcement.
-
Meta's AI Ad Platform Reportedly Causes Overspending and Poor Performance AIID ↗
Meta's automated ad platform "Advantage Plus" caused advertisers to exceed their daily ad budgets. The cost per impressions (CPMs) surged far above the usual. This incident, which persisted into April, affected small businesses with overspending and lack of transparency.
-
Deepfake Audio Falsely Attributes Controversial Remarks to Keir Starmer About the Rochdale Azhar Ali Crisis AIID ↗
A deepfake audio clip, falsely claiming to be Keir Starmer discussing the Rochdale byelection and Labour's withdrawl of support for Azhar Ali, circulated online, achieving over 250,000 views. Experts confirmed its inauthenticity, highlighting a significant misuse of AI in fabricating political content.
-
AI Romance Apps Reportedly Compromise User Privacy for Data Harvesting AIID ↗
AI-powered romantic chatbots, marketed for enhancing mental health, are found to exploit user privacy by harvesting sensitive personal information for data sharing and targeted ads, with inadequate security measures and consent protocols, according to research by the Mozilla Foundation.
-
Purported Deepfake Video Allegedly Claims Kyiv's Assassination Plan Against President Emmanuel Macron AIID ↗
A purported deepfake video reportedly claimed France 24 reported a Kyiv plot to assassinate French President Macron. This reported fake news was debunked by France 24, which confirmed the video was altered and did not air any such report.
-
Whisper Speech-to-Text AI Reportedly Found to Create Violent Hallucinations AIID ↗
Researchers at Cornell reportedly found that OpenAI's Whisper, a speech-to-text system, can hallucinate violent language and fabricated details, especially with long pauses in speech, such as from those with speech impairments. Analyzing 13,000 clips, they determined 1% contained harmful hallucinations. These errors pose risks in hiring, legal trials, and medical documentation. The study suggests improving model training to reduce these hallucinations for diverse speaking patterns.
-
Deepfake of Long-Deceased Suharto Circulating in Run-up to February 2024 Indonesian Elections AIID ↗
An AI-generated deepfake of Suharto, the deceased Indonesian dictator, was generated and circulated by the Golkar Party ahead of the February 2024 Indonesian elections. This video, which aimed to influence voter perceptions by invoking Suharto's legacy, sought to manipulate public opinion and misused deceased individuals' likenesses for political gain. The incident is another example of political deepfakes creating convincing misinformation.
-
Many Purported Political Deepfakes Reportedly Circulating in Run-up to 2024 Pakistani General Elections AIID ↗
During Pakistan's 2024 general elections, purportedly AI-generated deepfakes of a politically motivated nature were circulated. These deepfakes reportedly portrayed political figures in misleading contexts, spreading misinformation and aiming to influence voter perceptions and election outcomes.
-
Alleged Deepfake Audio of Imran Khan Calls for Election Boycott, Misleading Pakistan Voters AIID ↗
A purported deepfake audio clip, falsely attributed to Imran Khan urging a PTI (Pakistan Tehreek-e-Insaf) election boycott, circulated on social media on the eve of Pakistan's general elections. This sophisticated AI-generated misinformation aimed to mislead voters, highlighting the growing challenge of digital manipulation in political discourse.
-
Purported Russian-Linked Network Allegedly Used Deepfake of Maria Ressa on Facebook and Bing to Promote Cryptocurrency Scam Targeting Filipinos AIID ↗
A purportedly Russian-linked scam network allegedly circulated an AI-generated deepfake of journalist Maria Ressa on Facebook and Microsoft Bing. The video reportedly manipulated a 2022 interview to falsely depict Ressa endorsing cryptocurrency. Fraudulent websites allegedly impersonating Rappler and CNN Philippines reportedly amplified the scam. Investigators reportedly linked the operation to a fraudulent network targeting Filipino audiences. Meta and Microsoft removed the content, but similar
-
A Self-Driving Waymo Robotaxi Reportedly Collided with a Bicyclist AIID ↗
A Waymo robotaxi in San Francisco reportedly failed to detect a cyclist obscured by a truck, resulting in a collision with minor injuries, at 17th and Mississippi Streets in Potrero Hill. The incident reportedly demonstrated a vulnerability in autonomous vehicles' ability to safely navigate complex urban environments.
-
Facial Recognition System in Buenos Aires Triggers Police Checks Based on False Matches AIID ↗
Buenos Aires's facial recognition system mistakenly flagged innocent people as criminals, leading to wrongful stops and detentions. Judicial investigations indicate the technology may have been misused for unauthorized surveillance and data collection. Despite privacy risks, the system has been used widely without full disclosure of standards or safeguards,
-
Two Passengers Report Feeling Trapped in Waymo Car During Sensor Obstruction AIID ↗
During a nighttime Waymo ride in San Francisco, an individual reportedly blocked the sensors of an autonomous vehicle, leaving passengers Robert Moreno and his husband feeling trapped and unsure how to proceed. Waymo's support team reportedly advised the riders to stay inside the vehicle for safety. The individual reportedly left after a few minutes without further escalation.
Disclaimer: informational only, not legal advice. Verify every claim against the linked official sources and consult a qualified lawyer before acting.
Frequently asked questions
What is recorded in each incident?
What happened, when, who was involved as developer or deployer, who was harmed, the sectors and countries affected, and the domain and subdomain the harm falls under. Incidents are mirrored from the AI Incident Database and each links back to its source record.
Does an incident mean wrongdoing was proven?
No. These are reported incidents, not findings of liability. Entities named in a report are named as reported, allegations are marked as alleged, and nothing here is a legal determination.
How current is the incident data?
It syncs directly from the AI Incident Database several times a day, so a new incident usually appears within hours. Each record carries the date it was last synced, and the snapshot date is shown on the page.
Does this page include the full incident reports?
No. The report texts are outside the licence this data is shared under, so only the metadata and the classifications appear here. Every row links to the original incident record, which carries the reports themselves.