AIM: AI Incidents and Hazards Monitor

The article explicitly mentions AI-powered bots targeting APIs and causing application-layer attacks that disrupt services, which constitutes harm to digital infrastructure and communities relying on these services. The surge in attacks and reported security incidents indicates that harm is occurring, not just potential. The involvement of AI systems in these attacks and the resulting disruption aligns with the definition of an AI Incident, as the AI system's use has directly led to harm (disruption of critical digital infrastructure and services).[AI generated]

The Meta Alert System uses AI to analyze social media content for signs of suicidal intent, triggering alerts to police who then intervene. The AI system's outputs directly influenced real-world outcomes by enabling rapid rescue and medical treatment, preventing fatalities. The involvement of AI in detecting harmful content and facilitating timely intervention meets the criteria for an AI Incident, as it directly led to preventing injury or death. The article describes realized harm prevention rather than just potential risk, so it is not merely a hazard or complementary information.[AI generated]

The Grok chatbot is an AI system generating sexually explicit deepfake images without consent, which is a direct violation of rights and causes harm to individuals depicted. The investigations and court orders against xAI and Grok are responses to this harm. The involvement of AI in generating harmful content that has materialized harm fits the definition of an AI Incident. The political and trade policy discussions are complementary context but do not change the core classification.[AI generated]

The BlueCruise system is an AI system providing partial autonomous driving capabilities. The crashes caused fatalities, which constitute injury or harm to persons. The AI system's failure to act (no braking or steering) directly led to these harms. Therefore, this qualifies as an AI Incident because the AI system's use and malfunction directly caused harm to people.[AI generated]

The Baidu Apollo Go robotaxis are AI systems performing autonomous driving tasks. The 'system malfunction' caused these vehicles to stop unexpectedly, resulting in traffic jams and at least one crash, which constitutes harm to property and communities. The AI system's malfunction is the direct cause of this incident. Therefore, this qualifies as an AI Incident under the framework.[AI generated]

An AI system was explicitly used to assist in writing the review, and its outputs included unattributed material copied from another source, constituting a breach of intellectual property rights and journalistic ethics. This misuse of AI directly led to reputational harm to the journalist and the publication, as well as a breach of legal and ethical standards. Therefore, this qualifies as an AI Incident due to the realized harm involving violation of intellectual property rights and professional standards caused by the AI system's use.[AI generated]

The platforms' content moderation systems likely rely on AI to detect harmful content. The failure of these AI systems to accurately identify and remove child sexual exploitation and abuse material and terrorism content has resulted in the dissemination of such harmful content, which constitutes harm to communities and individuals. This meets the criteria for an AI Incident because the AI system's malfunction or inadequate performance has directly led to harm. The article details realized harm and regulatory actions taken in response, confirming the incident status rather than a mere hazard or complementary information.[AI generated]

The event involves the use of an AI system explicitly described as detecting harmful content related to digital sexual crimes and automating deletion requests. The AI system's use directly contributes to preventing harm to victims of sexual exploitation and abuse, which falls under harm to persons (a). Since the AI system is actively used to mitigate and respond to ongoing harm, this qualifies as an AI Incident rather than a hazard or complementary information. The article does not merely discuss potential risks or responses but reports on an operational AI system that has a direct role in harm prevention and victim protection.[AI generated]

The article involves AI systems in the form of advanced drones with likely autonomous capabilities used for military interception. While no harm has yet occurred, the production and export of such AI-enabled military drones could plausibly lead to AI incidents involving injury, disruption, or other harms. Therefore, this event fits the definition of an AI Hazard rather than an AI Incident or Complementary Information. It is not unrelated because the AI system's development and use are central to the event.[AI generated]

The use of AI to alter the campaign photo constitutes the use of an AI system. The resulting harm is indirect, as the AI-generated image misled voters and caused reputational damage and political controversy, which can be considered harm to the community or a violation of trust. Although the harm is non-physical and reputational, it is significant and directly linked to the AI system's use. Therefore, this qualifies as an AI Incident due to the realized harm caused by the AI-manipulated image in a political context.[AI generated]

The vulnerability involves AI agents within the Vertex AI platform, which qualifies as AI systems. The exploitation of default permission scoping to weaponize these AI agents directly leads to harm by enabling unauthorized data access and infrastructure compromise, which fits the criteria of an AI Incident under harm to property and critical infrastructure disruption. Therefore, this event is classified as an AI Incident.[AI generated]

The AI system is explicitly described as detecting abnormal behavior in real time and triggering alerts that enable security personnel to intervene promptly, preventing or reducing harm from crimes in unmanned stores. The article provides concrete examples where the AI system's detection led to immediate response and arrest, showing direct involvement in harm prevention. Therefore, this event qualifies as an AI Incident because the AI system's use has directly influenced the management of crime-related harms to property and community safety.[AI generated]

The project involves the development of a drone likely equipped with AI for autonomous or semi-autonomous operation, given the nature of military drones. Although no incident or harm has occurred yet, the mere development and potential deployment of AI-enabled military drones constitute an AI Hazard due to the credible risk of future harm such systems could cause. The article does not report any realized harm or incident, so it cannot be classified as an AI Incident. It is not merely complementary information since the focus is on the development of a potentially hazardous AI system, not on responses or updates to past incidents.[AI generated]

The article explicitly mentions the use of AI in scam calls and messages impersonating the IRS, which have led to actual harm including identity theft and financial fraud. The AI systems are used maliciously to generate convincing fake communications that deceive victims, causing direct harm to individuals' finances and personal data security. This fits the definition of an AI Incident because the AI system's use has directly led to harm to persons and communities through fraud and identity theft.[AI generated]

The article explicitly links the rise of AI-generated deepfake content to societal harm, including misinformation and potential damage to individuals and public discourse. The AI system's use in generating deepfakes has directly led to these harms, fulfilling the criteria for an AI Incident. The platforms' increased takedown efforts are responses to an ongoing incident rather than the main focus, so the article is not primarily about complementary information. The harm is realized and ongoing, not merely potential, so it is not an AI Hazard. Hence, the classification is AI Incident.[AI generated]

The cameras use AI to detect traffic violations and automatically issue fines, which directly affects individuals by imposing penalties. The AI system's outputs have led to concrete administrative actions (fines) and thus have caused realized harm. The event involves the use of an AI system, and the harm is direct and materialized, fulfilling the criteria for an AI Incident. The description does not merely discuss potential or future risks but actual enforcement actions taken based on AI detection, which excludes classification as a hazard or complementary information.[AI generated]

The event describes the use of facial recognition technology, which is an AI system, to process user data without consent, leading to a violation of privacy rights and breach of legal obligations. This constitutes harm under the category of violations of human rights or breach of applicable law protecting fundamental rights. Since the harm has already occurred and the settlement addresses this misuse, this qualifies as an AI Incident.[AI generated]

The article explicitly mentions AI systems being used maliciously to perpetrate scams that cause financial and emotional harm to individuals and businesses. The harms include fraud, identity theft, and exploitation of vulnerable groups, which fall under harm to persons and communities. The AI involvement is clear in the use of deepfake technology and AI-generated content to impersonate individuals and create fake companies. Since these harms are already occurring and the AI systems are pivotal in enabling these scams, the event meets the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]

The article explicitly mentions AI systems in the form of autonomous swarming drones with distributed control and multi-agent autonomy, which qualifies as AI systems. The event concerns the acquisition and planned integration of this technology into defense-related drone systems. No actual harm or incident is reported; the article focuses on business and technological development. However, the nature of the technology—autonomous swarming drones for battlefield use—implies a credible potential for future harm, such as injury, disruption, or rights violations, if deployed or misused. Therefore, the event fits the definition of an AI Hazard rather than an AI Incident or Complementary Information. It is not unrelated because the AI system and its potential impacts are central to the article.[AI generated]

The article involves AI systems in the context of military drones and their evolving capabilities, including potential future autonomous AI systems. However, it does not report any realized harm or incident caused by AI systems. The discussion is about ongoing use, strategic implications, and potential future developments, which aligns with the definition of an AI Hazard since it plausibly could lead to harm in the future. Yet, since no specific AI system malfunction or misuse causing harm is described, and the focus is more on broad analysis and future risks rather than a particular event, the classification as AI Hazard is appropriate. It is not Complementary Information because it is not updating or responding to a previously reported incident, nor is it unrelated as it clearly involves AI-related military technology and its implications.[AI generated]