AIM: AI Incidents and Hazards Monitor

The autonomous bus is an AI system as it performs autonomous navigation and decision-making. The collision with the tramway is a direct consequence of the AI system's operation, constituting harm to property and potential risk to passenger safety. Although no injuries occurred, the incident itself is a realized harm event involving an AI system. Hence, it meets the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]

The forward collision-avoidance system is an AI system that processes camera input to make real-time decisions about braking to prevent collisions. The software bug causing premature activation of braking has directly resulted in crashes and injuries, fulfilling the criteria for an AI Incident due to harm to persons. The recall and software update are responses to this malfunction. Therefore, this event qualifies as an AI Incident.[AI generated]

An AI system is explicitly involved as it autonomously pilots aircraft and communicates with control towers. The event stems from the use and development of this AI system. Although no incident or harm has yet occurred, the article discusses potential safety risks and labor concerns, indicating plausible future harms. Hence, this qualifies as an AI Hazard rather than an AI Incident or Complementary Information, since the tests are ongoing and no harm has materialized yet.[AI generated]

The article explicitly mentions AI systems integrated into autonomous drones, including AI inference chips, image recognition algorithms, and secure communication modules. The partnership aims to develop these technologies for both commercial and military uses, which inherently carry risks of misuse or malfunction leading to harm. No actual harm or incident is reported, so it is not an AI Incident. The focus is on the development and strategic positioning of AI-enabled drone technologies with plausible future risks, fitting the definition of an AI Hazard. It is not merely complementary information because the article centers on the creation of potentially hazardous AI systems rather than updates or responses to existing incidents.[AI generated]

The article clearly involves AI systems in the form of AI data centers and infrastructure, which are being affected by war and geopolitical risks. The attacks on data centers and the increased security concerns indicate a credible risk that these AI systems could be disrupted, potentially leading to harm such as disruption of critical infrastructure. However, the article does not report any actual harm or incident caused by AI system malfunction or misuse. The focus is on the plausible future harm to AI infrastructure and the strategic responses to these risks. Therefore, this event fits the definition of an AI Hazard, as it plausibly could lead to an AI Incident but no incident has yet occurred.[AI generated]

The event involves an AI system (AI coding assistants Claude and Cursor) being hijacked by malware to perform malicious actions that result in theft of crypto assets and sensitive credentials. This constitutes direct harm to property and security, fulfilling the criteria for an AI Incident. The malware's use of AI to facilitate the attack and the realized harm from stolen data and crypto assets confirm this classification. The event is not merely a potential risk or a general update but a concrete incident involving AI system misuse causing harm.[AI generated]

The digital triage apps qualify as AI systems because they perform automated decision-making to recommend medical triage outcomes. The event describes the use and mandated adoption of these AI systems despite insufficient evidence of safety, leading to concerns about patient harm and compromised clinical judgment. The financial penalties imposed on doctors who refuse to use these apps create indirect pressure that could lead to harm. This fits the definition of an AI Incident, as the AI system's use has directly or indirectly led to potential or actual harm to patient health (harm category a). The article reports ongoing harm and risk, not just potential future harm, so it is not merely an AI Hazard. It is not Complementary Information because the article focuses on the conflict and risks arising from the AI system's use, not on responses or updates to prior incidents. It is not Unrelated because the AI system and its impact are central to the event.[AI generated]

The article explicitly mentions AI-powered delivery robots operating autonomously in public spaces, causing physical injuries (cyclist with head injury and broken collarbone), property damage (broken bus shelter glass), and social disruption (blocking sidewalks, interfering with pedestrian traffic). These are direct harms linked to the use of AI systems. The presence of injuries and property damage confirms realized harm rather than just potential harm. Hence, this qualifies as an AI Incident under the OECD framework, as the AI system's use has directly or indirectly led to harm to persons and communities.[AI generated]

The event involves AI system development and use (training AI with video data from private homes), which raises plausible privacy and human rights concerns. However, the article does not describe any actual harm or violation that has occurred, only a controversy and criticism. Therefore, it fits the definition of an AI Hazard, as the AI system's use could plausibly lead to harm (privacy violations, human rights breaches) but no incident has been reported yet.[AI generated]

The article explicitly states that AI was used to mimic voices and images of government officials to carry out fraud, which is a direct harm to individuals and communities through deception and financial loss. The AI system's use in this criminal activity led to realized harm, meeting the criteria for an AI Incident. The law enforcement response and arrests confirm the harm has occurred and is being addressed.[AI generated]

The article centers on a research position aimed at studying and mitigating potential future risks from AI systems capable of recursive self-improvement. While it involves AI systems and their development, the event does not describe any realized harm or incident caused by AI. Instead, it discusses plausible future threats and the proactive measures to address them. Therefore, this qualifies as an AI Hazard, as it concerns credible potential risks that could arise from AI development but have not yet materialized.[AI generated]

The event involves an AI system (Claude Mythos) whose capabilities to penetrate cybersecurity defenses could plausibly lead to disruption of critical infrastructure (European banks). The ECB's urgent response and the discussion of rapid patching processes indicate recognition of a credible threat. Since no actual harm has been reported yet, but the threat is imminent and plausible, this qualifies as an AI Hazard rather than an AI Incident. The article focuses on the potential for harm and the need for urgent action to prevent it.[AI generated]

The event involves AI systems (autonomous AI agents executing transactions) and discusses a credible systemic risk stemming from their dependence on a single stablecoin issuer. The risk of a regulatory challenge or de-peg event causing widespread disruption is a plausible future harm linked to the AI system's use. Since no actual harm has been reported yet, but the potential for significant disruption exists, this qualifies as an AI Hazard rather than an AI Incident. The article does not focus on responses or updates but on the emerging risk itself.[AI generated]

The article explicitly mentions AI chatbots and AI-generated images influencing patients' expectations for plastic surgery, leading to requests for procedures that are unrealistic or potentially harmful. The AI systems' outputs are directly linked to patient demands that could cause physical harm, psychological harm, and financial harm. This fits the definition of an AI Incident because the AI system's use has indirectly led to harm to persons (patients) through unrealistic expectations and potentially harmful surgeries. The involvement is through the use of AI systems (chatbots and image generators) shaping patient perceptions and decisions, fulfilling the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]

The event involves AI systems explicitly—voice AI models used in smart speakers and phones. The hackers' adversarial audio exploits the AI system's malfunction or vulnerability to cause unauthorized actions, directly leading to harm in terms of privacy breaches and potential access to sensitive information. The harm is realized or highly plausible given the demonstrated attack method. Therefore, this qualifies as an AI Incident due to direct harm caused by the AI system's exploitation.[AI generated]

The event involves the use of AI synthesis technology to create fake live streaming backgrounds that misrepresent product origins, leading consumers to make purchasing decisions based on false information. This constitutes a violation of consumer rights and legal protections against false advertising, which are harms to individuals and communities. The AI system's role in fabricating these deceptive scenes is pivotal and directly linked to the harm caused. Therefore, this qualifies as an AI Incident rather than a hazard or complementary information.[AI generated]

The article involves AI systems indirectly through hyperscale AI data centres, which are critical infrastructure supporting AI workloads. However, the reported issue is about the government's failure to account for their emissions in climate impact assessments, representing a potential future harm (environmental impact) rather than a realized harm. There is no direct or indirect harm caused by AI system malfunction or misuse described. Therefore, this qualifies as an AI Hazard, as the development and operation of these AI data centres could plausibly lead to significant environmental harm if unaddressed.[AI generated]

The article explicitly mentions fake AI-generated images being created and spread, which is a direct misuse of AI systems generating harmful content. The harm here is a violation of privacy, which falls under the category of violations of human rights or breach of obligations intended to protect fundamental rights. The actor's response and legal considerations further confirm the recognition of harm caused by the AI system's outputs. Therefore, this event qualifies as an AI Incident.[AI generated]

The article explicitly states that AI-based tools are being used by cybercriminals to automate reconnaissance, detect vulnerabilities, and accelerate attacks, leading to a drastic reduction in exploitation time and a large increase in ransomware victims. These harms include data theft, disruption, and financial losses, which fall under harm to property and communities. The AI systems' malicious use is a direct contributing factor to these harms. Hence, the event meets the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]

The video was created using AI to generate false images and speech, which directly caused harm by misleading the public and causing internal political crisis and distress to the individual portrayed. The unauthorized use of a person's likeness and fabricated statements constitutes a violation of rights and harm to reputation, fitting the definition of an AI Incident. The AI system's use in producing the deepfake video is central to the harm realized.[AI generated]