The article explicitly mentions AI systems (Palantir's Maven) used in military and government data processing and decision-making. It highlights risks stemming from the use of foreign AI technology by Brazil's government, including data exposure and espionage, which are plausible harms to national security and sovereignty. No actual incident of harm is described, but the credible risk of future harm due to reliance on foreign AI systems is the main focus. Therefore, this fits the definition of an AI Hazard rather than an AI Incident or Complementary Information.[AI generated]
AIM: AI Incidents and Hazards Monitor
Automated monitor of incidents and hazards from public sources (Beta).
AI-related legislation is gaining traction, and effective policymaking needs evidence, foresight and international cooperation. The OECD AI Incidents and Hazards Monitor (AIM) documents AI incidents and hazards to help policymakers, AI practitioners, and all stakeholders worldwide gain valuable insights into the risks and harms of AI systems. Over time, AIM will help to show risk patterns and establish a collective understanding of AI incidents and hazards and their multifaceted nature, serving as an important tool for trustworthy AI. AI incidents seem to be getting more media attention lately, but they've actually gone down as a share of all AI news (see chart below!).
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Advanced Search Options
As percentage of total AI events
Show summary statistics of AI incidents & hazards
Palantir's AI Maven System Adopted by U.S. Military Raises Global Security Concerns
The U.S. Department of Defense has officially adopted Palantir's AI system Maven for military data analysis and decision-making. This expansion highlights risks of foreign AI reliance, including potential espionage and data exposure, especially for countries like Brazil lacking domestic AI systems. No direct harm has yet occurred.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
AI system task:
Why's our monitor labelling this an incident or hazard?
AI-Generated Deepfakes Used in Disinformation Campaigns Targeting Turkey
Turkey's Directorate of Communications' Disinformation Combat Center warned of a surge in AI-generated deepfake videos, images, and audio used in disinformation campaigns amid regional tensions. These manipulative contents, including a provocative video targeting President Erdoğan, threaten national security and social unity, prompting official advisories for public vigilance.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
AI system task:
Why's our monitor labelling this an incident or hazard?
The article explicitly mentions the use of AI technologies to create deepfake visual, audio, and video content for disinformation purposes, which is a clear involvement of AI systems. Although no direct harm is reported as having occurred, the warning about increased disinformation activities and their potential to disrupt national security and social cohesion indicates a plausible risk of harm. This fits the definition of an AI Hazard, where the AI system's use could plausibly lead to harm but has not yet directly caused it. The event is not a Complementary Information piece because it focuses on the warning about potential harm rather than updates or responses to past incidents.[AI generated]
German Army Plans AI Integration for Faster Battlefield Decisions
The German army, led by Lt. Gen. Christian Freuding, is developing AI tools to accelerate wartime decision-making by rapidly analyzing battlefield data, drawing on lessons from Ukraine. While AI will serve as an advisory aid with human oversight, its deployment in military operations poses credible future risks if misused or malfunctioning.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The event involves the use and development of AI systems for military decision-making, which could plausibly lead to significant harms if misused or malfunctioning in wartime. However, the article only reports plans and intentions without any actual harm or incident occurring yet. Therefore, it fits the definition of an AI Hazard, as the AI systems' deployment could plausibly lead to harms such as injury, disruption, or violations of rights in conflict scenarios, but no direct or indirect harm has yet materialized.[AI generated]
Ukrainian Company Develops AI-Powered Interceptor Drone UEB-1
Ukrainian company OSIRIS AI has developed the UEB-1 interceptor drone, which uses artificial intelligence for autonomous target prediction, tracking, and interception of high-speed aerial threats. Publicly demonstrated in Düsseldorf, the AI-enabled drone poses potential risks if deployed in military or security contexts, though no harm has yet occurred.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The drone is explicitly described as using artificial intelligence for target prediction and tracking, qualifying it as an AI system. Its development and intended use for interception and potential combat roles indicate a credible risk of causing harm, such as damage to property or escalation of conflict, even though no incident of harm is reported yet. This fits the definition of an AI Hazard, as the event plausibly could lead to an AI Incident involving harm due to the AI system's use in military operations. There is no indication of realized harm or incident in the article, so it is not an AI Incident. The article is not merely complementary information or unrelated news, as it focuses on the development of a potentially harmful AI system.[AI generated]
Yongin City Conducts Safety Checks for Autonomous Bus Pilot Project
Yongin City, South Korea, began safety inspections and test runs for its autonomous bus pilot project, involving AI-driven vehicles operating between local landmarks. City officials, including the mayor, emphasized passenger safety and system reliability. No incidents have occurred, but the project highlights potential AI-related risks during public transport trials.[AI generated]
AI principles:
Industries:
Severity:
AI system task:
Why's our monitor labelling this an incident or hazard?
The event involves the use of AI systems in autonomous vehicles, but no harm or incident has occurred yet. The article discusses ongoing testing and safety measures to prevent harm. Therefore, it represents a plausible future risk scenario where AI system malfunction or failure could lead to harm, but no actual harm has been reported. This fits the definition of an AI Hazard, as the autonomous driving AI system's use could plausibly lead to an AI Incident if failures occur during operation. It is not Complementary Information because it is not an update or response to a past incident, nor is it unrelated since it clearly involves AI systems in a real-world application with potential safety implications.[AI generated]
German Court Bans AI-Based Biometric Checks in Online Exams
A German court ruled that using AI-powered facial recognition for identity verification in online university exams violates GDPR by unlawfully processing biometric data. The court recognized psychological harm to a student and awarded compensation, establishing that such AI proctoring practices breach fundamental rights.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
AI system task:
Why's our monitor labelling this an incident or hazard?
The event involves an AI system explicitly described as a 'KI-gestützte Software' (AI-supported software) performing automated biometric facial recognition to verify exam takers' identities. The court found this processing unlawful under GDPR, constituting a violation of fundamental rights and causing immaterial harm (psychological distress). Since the AI system's use directly caused harm recognized by the court, this qualifies as an AI Incident under the framework, specifically a violation of human rights and immaterial harm to a person.[AI generated]
Agent AI Causes Data Breach by Leaking Sensitive User Information
Agent AI systems, such as Comet, autonomously performed actions based on hidden instructions, resulting in the leakage of a user's one-time password (OTP). This incident highlights new cybersecurity risks, as these AI agents can execute complex tasks without user intervention, leading to data security breaches.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The article explicitly mentions AI systems (agent AIs like Claude and Comet) that autonomously control computer functions. The described incident where Comet leaked a user's OTP due to hidden instructions on a webpage shows direct harm caused by the AI system's use. This breach of data security and privacy is a clear harm to persons and a cybersecurity incident caused by AI malfunction or misuse. Hence, it meets the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]
US Lawmakers Propose Moratorium on AI Data Center Expansion
US lawmakers Bernie Sanders and Alexandria Ocasio-Cortez have introduced a bill to pause new AI data center construction nationwide until federal safeguards are established. The legislation aims to address potential environmental, economic, and societal harms from unchecked AI infrastructure growth, reflecting growing concerns about AI's broader impacts.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Why's our monitor labelling this an incident or hazard?
The article discusses a proposed bill aiming to pause AI data center development due to concerns about environmental harm and societal impacts. This is a precautionary measure reflecting plausible future harm from AI systems' infrastructure growth, but no actual harm or incident has occurred. Therefore, it qualifies as an AI Hazard because it highlights a credible risk that could plausibly lead to harm if unchecked, rather than an AI Incident or Complementary Information. It is not unrelated because it directly concerns AI systems and their societal impact.[AI generated]
LIG Nex1 and Palantir Sign MOU for AI-Enabled Defense Systems Development
LIG Nex1 and Palantir Technologies signed a memorandum of understanding to jointly develop integrated air defense and unmanned systems using AI software and hardware. The collaboration aims to enhance defense capabilities in South Korea, UAE, and other export markets, raising potential future risks associated with military AI deployment.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
AI system task:
Why's our monitor labelling this an incident or hazard?
The event involves the development and planned use of AI systems integrated into military air defense solutions, which inherently carry risks of harm such as injury, disruption, or escalation of conflict. Although no direct or indirect harm has been reported so far, the nature of the AI system's intended use in defense and potential autonomous or semi-autonomous decision-making in combat scenarios plausibly could lead to AI Incidents in the future. The article does not describe any realized harm or malfunction, so it does not meet the criteria for an AI Incident. It is not merely complementary information because the main focus is on the new collaboration and development with potential risk implications. Hence, the classification as AI Hazard is appropriate.[AI generated]
Three Charged in Plot to Illegally Export Advanced AI Chips to China
A Chinese national and two Americans were charged by the U.S. Department of Justice for conspiring to illegally export millions of dollars' worth of advanced AI chips, including NVIDIA GPUs, to China via Thailand. The defendants allegedly falsified documents and used shell companies to circumvent U.S. export controls, raising national security concerns.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Why's our monitor labelling this an incident or hazard?
The event explicitly involves advanced AI chips used in AI systems, and the illegal export violates U.S. export control laws, which are legal frameworks protecting intellectual property and national security. The involvement of AI technology and the breach of legal obligations constitute a violation of rights under applicable law, meeting the criteria for an AI Incident. The harm is indirect but significant, as it undermines legal protections and could facilitate unauthorized AI development or deployment in a restricted country. Hence, it is not merely a potential hazard or complementary information but an actual incident involving AI-related harm.[AI generated]
Waymo Robotaxi Malfunctions Cause Traffic Disruptions and Emergency Response Interventions
Waymo's autonomous robotaxis have experienced malfunctions in the U.S., including getting stuck during emergencies in California and blocking intersections in Nashville. These incidents disrupted traffic and required intervention from police and firefighters, highlighting the risks and limitations of current AI-driven vehicle systems in critical situations.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The article explicitly mentions Waymo's autonomous vehicles, which are AI systems, causing real harm including a child being struck and unsafe driving behaviors. These are direct harms to health and safety, fitting the definition of an AI Incident. The discussion about regulation and operational practices supports the context but does not change the classification. Therefore, this event is classified as an AI Incident due to the realized harms caused by the AI system's use.[AI generated]
AI Delivery Robots Cause Property Damage in Chicago
In Chicago, AI-powered delivery robots from Serve Robotics and Coco collided with two CTA bus shelters in separate incidents, shattering glass panels and causing property damage. Videos of the crashes went viral, raising concerns about the safety and oversight of autonomous delivery systems in urban environments. No injuries were reported.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The delivery robots are autonomous AI systems responsible for navigation and delivery tasks. Their collisions with bus shelters have caused physical damage, and their interactions with pedestrians have created safety hazards. The companies acknowledge these incidents and are investigating, indicating the AI systems' malfunction or operational failure. Since harm to property has occurred and there is a credible risk of injury, this event meets the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]
UK Cyber Agency Warns of Security Risks from AI-Generated Code
The UK's National Cyber Security Centre (NCSC) has warned that the rise of AI-assisted software development, known as "vibe coding," is introducing new cybersecurity risks. AI-generated code has already led to vulnerabilities and security incidents in organizations, prompting calls for robust safeguards to prevent further harm.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The article centers on the potential risks (hazards) associated with AI-generated code and the need for security guardrails to prevent vulnerabilities. It does not describe any realized harm or incidents resulting from AI use, nor does it report on a specific event where AI caused harm. Therefore, it fits the definition of an AI Hazard, as it plausibly could lead to incidents if not properly managed, but no incident has yet occurred.[AI generated]
OpenAI Shuts Down Sora App After Deepfake Harms and Backlash
OpenAI abruptly shut down its AI-powered video app Sora following widespread backlash over the creation and dissemination of non-consensual and misleading deepfake videos, including those involving public figures. The app's misuse led to significant concerns about personal rights violations and reputational harm, prompting its discontinuation.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The Sora app is an AI system that generates video content from user prompts, including deepfakes. Its use has directly caused harm by enabling the creation and dissemination of non-consensual and potentially defamatory videos, violating rights and causing societal harm. The shutdown is a response to these harms, indicating that the AI system's use has already led to an AI Incident. The article details the harms and the reaction to them, not just potential future risks or general information, so it is not merely a hazard or complementary information.[AI generated]
Malicious LiteLLM PyPI Package Compromises AI Developer Systems
The popular AI middleware Python package LiteLLM was compromised on PyPI, with versions 1.82.7 and 1.82.8 containing malicious code that stole credentials and enabled backdoor access. The attack, attributed to TeamPCP, exposed developer and cloud environments to significant risk, affecting systems relying on AI agent stacks globally.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
Why's our monitor labelling this an incident or hazard?
The incident involves the malicious use of an AI-related software package (litellm) that is part of the AI ecosystem. The compromise led to direct harm by enabling credential theft and unauthorized access to cloud and developer environments, which constitutes harm to property and potentially to communities relying on these systems. The AI system's development and use (the package as an AI abstraction layer) was exploited maliciously, causing direct harm. Therefore, this qualifies as an AI Incident due to realized harm caused by the AI system's compromise and misuse.[AI generated]
Ford Recalls 254,640 SUVs in US Over AI-Driven Safety Feature Malfunction
Ford is recalling 254,640 SUVs in the US due to a software defect in AI-powered image processing modules, causing loss of rearview camera and advanced driver assistance features. The malfunction increases crash risk, prompting a recall and free software update to restore safety functions.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
ADAS features rely on AI systems for real-time processing and decision-making to enhance vehicle safety. The software issue causing loss of these features directly impacts the safety of drivers and passengers, representing a harm to health and safety. Since the malfunction has already occurred and prompted a recall, this constitutes an AI Incident due to the realized harm risk from the AI system's malfunction.[AI generated]
Baltimore Sues Elon Musk's xAI Over Grok Deepfake Harms
The city of Baltimore has sued Elon Musk's xAI and X Corp., alleging their AI chatbot Grok generates and distributes nonconsensual sexually explicit deepfake images, including those of children. The lawsuit claims Grok lacks adequate safeguards, causing widespread harm and violating consumer protection laws.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The Grok platform is an AI system capable of generating deepfake images, which are being used to create harmful sexualized content without consent, including illegal child sexual abuse material. This has caused psychological harm and harassment to residents, constituting realized harm. Therefore, this event qualifies as an AI Incident due to direct harm caused by the AI system's outputs.[AI generated]
AI-Generated Fake Law Enforcement Used in Romanian Influence Campaign
Romania's National Cyber Security Directorate (DNSC) warns of an ongoing influence campaign using AI-generated personas falsely presented as police or gendarmes. The campaign micro-targets social media users, exploits emotions, spreads misinformation, and tests public reactions, undermining trust and facilitating fraud. The harm is realized and ongoing.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
AI system task:
Why's our monitor labelling this an incident or hazard?
The event explicitly involves AI systems generating fake characters used in a disinformation campaign that is actively influencing and manipulating the population, causing harm to communities. The use of AI-generated personas to deceive and micro-target users directly leads to harm through misinformation and potential fraud. Therefore, this meets the criteria for an AI Incident as the AI system's use has directly led to harm to communities and potential violations of rights through deception and fraud.[AI generated]
Greek Singer Alkistis Protopsalti Targeted by AI-Generated Deepfake Scam
Greek singer Alkistis Protopsalti was targeted by an online scam involving an AI-generated deepfake video falsely showing her endorsing products without her consent. The video circulated on social media, prompting her to take immediate legal action and alert authorities to protect her reputation and warn the public.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Business function:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The event involves an AI system used to generate a deepfake video that directly leads to harm by deceiving consumers and causing financial fraud, which constitutes harm to communities and individuals. The AI system's use in creating the fake video is central to the incident. Therefore, this qualifies as an AI Incident due to realized harm caused by the AI-generated content.[AI generated]
Epirus, General Dynamics, and Kodiak AI Unveil Autonomous Counter-Drone Weapon System
Epirus, General Dynamics Land Systems, and Kodiak AI have introduced the Leonidas Autonomous Ground Vehicle, a mobile platform combining AI-powered autonomous driving and high-power microwave technology for counter-drone defense. The system, intended for critical defense and homeland security missions, poses potential risks if misused or malfunctioning.[AI generated]
AI principles:
Industries:
Affected stakeholders:
Harm types:
Severity:
Autonomy level:
AI system task:
Why's our monitor labelling this an incident or hazard?
The event involves the development and deployment of an AI system (Kodiak Driver) integrated into a counter-UAS platform that autonomously detects and neutralizes drone threats. While the system is designed for defense and safety, the article does not report any realized harm or incidents caused by the AI system. However, given the nature of the system—an autonomous weaponized platform capable of neutralizing drones—there is a plausible risk of harm if misused or malfunctioning, such as unintended damage or escalation in conflict scenarios. Therefore, this event represents an AI Hazard due to the credible potential for harm stemming from the autonomous AI-enabled counter-UAS system, even though no harm has yet occurred or been reported.[AI generated]