AIM: AI Incidents and Hazards Monitor

No actual harm has yet occurred, but the coalition’s analysis shows that continued AI‐led expansion of data centers could plausibly lead to large‐scale environmental damage (increased greenhouse gas emissions, strain on renewable supplies). Therefore, this is an AI hazard scenario, not a realized incident or merely complementary context.[AI generated]

The article explicitly mentions an AI system controlling traffic lights in real time to optimize traffic flow and reduce congestion and accidents. The system's deployment has resulted in a 33.8% reduction in traffic accident rates and improved traffic speeds, indicating a direct link between AI use and harm reduction (less injury and improved safety). Since the AI system's use has directly led to a measurable decrease in harm (accidents), this event fits the definition of an AI Incident. The harm here is positive (reduction of harm), but the framework includes injury or harm to health, so the AI system's role in reducing such harm is relevant. This is not a hazard or complementary information, as the AI system is actively in use and causing real-world effects.[AI generated]

Tesla's Autopilot is an AI system enabling partially automated driving. The article reports multiple crashes involving this technology, including a fatal accident, which are under federal investigation. These investigations and recalls are safety measures addressing harms caused by the AI system's malfunction or limitations. The article also highlights the risk that political influence could weaken these safety measures, increasing future harm. Since actual harm has occurred due to the AI system's use, this qualifies as an AI Incident. The discussion of potential weakening of oversight is relevant but secondary to the realized harms.[AI generated]

The event involves AI systems (conversational bots, deepfakes) used in dating apps that have directly led to harms including emotional manipulation, deception, and phishing attacks on users. These constitute harm to people and communities, fulfilling the criteria for an AI Incident. The article describes realized harms, not just potential risks, and the AI's role is pivotal in causing these harms. Therefore, this is classified as an AI Incident.[AI generated]

The article explicitly mentions AI-generated images being used in scams that have caused customers to lose significant amounts of money. This constitutes direct harm to individuals' property (financial loss) caused by the use of an AI system. Therefore, this qualifies as an AI Incident under the framework, as the AI system's use has directly led to harm.[AI generated]

The event centers on a new hack—indirect and delayed prompt injections—that can override Gemini’s defenses and permanently implant malicious instructions or false user data in the chatbot’s memory. Although demonstrated only in controlled research, it highlights a clear, plausible pathway to user misinformation, persistent behavioral manipulation, and data theft. Because this exploit poses a potential future harm rather than reporting actual widespread damage, it qualifies as an AI Hazard.[AI generated]

The report describes a real complaint filed to the French data protection authority (CNIL) alleging that an AI system (Le Chat) actively exploited personal data without proper consent, constituting a violation of users’ fundamental privacy rights under applicable law. This is a concrete AI‐driven harm (GDPR violation), so it qualifies as an AI Incident.[AI generated]

The article involves AI in the context of its potential societal and economic impacts, highlighting plausible future harms such as mass unemployment, social conflict, and environmental degradation linked to AI-driven productivity increases. However, it does not report any actual incident or harm caused by an AI system at present, nor does it describe a specific AI system malfunction or misuse. The focus is on the need for global regulation to prevent these potential harms, making this an AI Hazard rather than an AI Incident or Complementary Information.[AI generated]

The Vehicle-Mounted Counter-Drone System clearly involves AI systems, as it uses automatic detection, classification, and neutralization of drones through advanced sensors and integrated technologies. The event focuses on the system's unveiling and its capabilities, with no indication of any harm or incident resulting from its use or malfunction. Since the system is designed to counter drone threats, it could plausibly lead to harm if it malfunctions or is misused (e.g., wrongful targeting or escalation), but no such incident is described. Hence, it fits the definition of an AI Hazard rather than an AI Incident or Complementary Information. It is not unrelated because it involves AI-enabled defense technology with potential implications for harm.[AI generated]

This is a direct harm incident caused by the AI system’s malfunction. Tesla’s FSD, an autonomous driving system, failed to detect or respond to a lane ending, leading to property damage and potential risk to pedestrians. Under the framework, a realized harm stemming from an AI system’s malfunction qualifies as an AI Incident.[AI generated]

AI‐powered summarisation systems were directly used to generate news summaries and were found to produce incorrect facts, altered quotations, and blurred fact/opinion distinctions. Even though this is framed as a broad study rather than a single isolated event, it documents realized harms—misinformation and distortion of trusted sources—caused by deployed AI systems. Therefore, it qualifies as an AI Incident.[AI generated]

TikTok's content recommendation algorithm is an AI system that influences what content users see. The lawsuit alleges that this AI system deliberately targeted children with harmful content, leading to their deaths. This constitutes direct harm to persons caused by the use of an AI system, fulfilling the criteria for an AI Incident. The harm is realized and severe (death), and the AI system's role is pivotal as per the allegations. Therefore, this event is classified as an AI Incident.[AI generated]

The article details a real harm caused by the illegal sharing of private sexual content, which is a violation of rights and personal privacy. The mention of the deep fake porn industry implies the involvement of AI systems in creating or distributing manipulated sexual content. The ongoing widespread distribution of the video and the investigation into the sources and advertising around it indicate direct harm linked to AI-enabled technologies. Hence, this is an AI Incident as the AI system's use or misuse has directly led to harm to an individual and potentially others in similar situations.[AI generated]

The described robots use AI-powered targeting, autonomous navigation, and grenade‐avoidance capabilities, and they have directly caused lethal harm to Russian soldiers. This is a materialized harm (injury and death) caused by the use of AI systems in warfare, fitting the definition of an AI Incident.[AI generated]

The article explicitly mentions generative AI systems cloning voices of dubbing actors without their consent, which is a direct use of AI technology. This unauthorized use leads to harm by violating actors' rights and reducing their employment opportunities, fulfilling the criteria for an AI Incident under violations of human rights and labor rights. The harm is realized, not just potential, as actors report reduced work and ethical concerns. Hence, this is an AI Incident rather than a hazard or complementary information.[AI generated]

This is a direct misuse of an AI system (deepfake generation) to harass, threaten, and inflict emotional and reputational harm on Hannah and other women. The harm has materialized (non-consensual sexual content, threats, doxxing), constituting violations of personal and human rights. Therefore, it meets the criteria for an AI Incident.[AI generated]

The article describes actual harm (financial fraud) directly enabled by an AI system (voice-cloning technology). The AI’s misuse led to property harm, meeting the criteria for an AI Incident.[AI generated]

The article clearly involves an AI system, as Amorphous is an autonomy software platform coordinating large numbers of autonomous systems. The event concerns the development and use of this AI system for military swarm control. Although the software is not reported to have caused any harm yet, its deployment in military contexts with autonomous capabilities could plausibly lead to harms such as injury, disruption, or violations of rights. Therefore, this event fits the definition of an AI Hazard, as it describes a credible potential for future harm stemming from the AI system's use, but no realized harm or incident is described.[AI generated]

The article explicitly mentions AI systems (computer vision) being developed for defense purposes, which inherently carry risks of harm due to their application in warfare. Although no incident has occurred yet, the partnership's focus on military AI systems plausibly leads to potential harms such as injury or disruption, qualifying this as an AI Hazard rather than an Incident or Complementary Information.[AI generated]