AIM: AI Incidents and Hazards Monitor

The article does not report a realized harm or incident caused by AI notetakers but focuses on the potential risks and uncertainties related to their use, especially regarding data privacy and confidentiality. These concerns indicate a credible risk that the AI systems' development and use could plausibly lead to harms such as violations of rights and breaches of confidentiality. Therefore, the event fits the definition of an AI Hazard, as it describes circumstances where AI systems could plausibly lead to harm, but no direct or indirect harm has yet been reported.[AI generated]

The article explicitly states that AI-generated manipulated images are being disseminated, causing false narratives and reputational damage to Shakira. This constitutes harm to a person and communities through misinformation and unauthorized use of her image. The AI system's role is pivotal in creating these false images, making this an AI Incident under the definitions provided.[AI generated]

Character.AI is an AI system (a generative conversational AI chatbot). The event involves the use of this AI system and its failure to adequately protect minors and comply with data protection laws, which are legal obligations protecting fundamental rights. The imposition of a fine by the privacy authority confirms that harm (legal violations) has occurred. This fits the definition of an AI Incident under category (c) violations of human rights or breach of obligations under applicable law intended to protect fundamental rights. The event is not merely a potential risk or a complementary update but a concrete sanction for realized legal violations linked to the AI system's use. Hence, the classification is AI Incident.[AI generated]

The event involves the use of an AI system (Claude Fable 5) that purportedly replaced most of the software development workforce, leading to layoffs. This constitutes a violation of labor rights and harm to employment, which fits the definition of an AI Incident. Although the information is unverified and anonymous, the description clearly links the AI system's use to realized harm (job losses). Therefore, it qualifies as an AI Incident rather than a hazard or complementary information.[AI generated]

The event involves an AI system (facial recognition and risk scoring) used in surveillance and profiling, which directly led to harm through privacy violations and exposure of sensitive personal data after a data breach. The involvement of AI in tracking and labeling individuals with risk scores and sensitive attributes, combined with the leak of this data, constitutes a violation of human rights and privacy. This meets the criteria for an AI Incident as the AI system's use directly led to harm.[AI generated]

The event involves an AI system providing treatment sensitivity classifications that were inaccurate and ineffective. The clinicians' reliance on these faulty AI recommendations, despite contradictory patient recovery data, indicates that the AI system's use indirectly led to potential harm to patients' health. This fits the definition of an AI Incident because the AI system's use has directly or indirectly led to harm to persons (patients) through erroneous treatment decisions. The study's findings emphasize the real-world risk of harm from AI misuse or overreliance in healthcare, not just a hypothetical hazard or complementary information.[AI generated]

The article explicitly states that AI tools were used to create fake online identities, which were instrumental in the scam that defrauded elderly Americans of over $8 million. This use of AI directly led to significant harm to individuals, fulfilling the criteria for an AI Incident. The harm is realized (not just potential), involving violations of rights and financial harm to vulnerable groups. Hence, the event is classified as an AI Incident.[AI generated]

The article explicitly mentions that AI models have increased the ability to analyze and exploit old code, leading to a recent exploit causing a $4.7 million loss. This is a direct link between AI use and realized harm (financial loss) in a critical infrastructure-like system (blockchain bridges and smart contracts). The AI involvement is in the use phase, assisting attackers in exploiting vulnerabilities. The harm is financial loss to users and ecosystem disruption, which fits under harm to property and communities. Thus, the event meets the criteria for an AI Incident due to AI's pivotal role in enabling the exploit and consequent harm.[AI generated]

The article explicitly involves AI systems used in healthcare decision-making, which are central to the allegations. The lawsuit claims that safety and privacy standards were compromised, which could lead to faulty medical decisions (harm to health) and breaches of patient privacy (violation of rights). These harms fall under the AI Incident definition, as they are directly or indirectly caused by the development and use of AI systems. The presence of a lawsuit and specific allegations of misconduct and harm further support classification as an AI Incident rather than a hazard or complementary information. The event is not merely a potential risk but involves claims of actual harm or misconduct related to AI research and deployment.[AI generated]

The article explicitly describes an AI system (Palantir's Maven Smart System and other AI-enabled sensors, drones, and autonomous weapons) being developed and integrated into NATO's defense strategy to detect and engage enemy forces automatically. The AI's role is pivotal in decision-making and target engagement, which could lead to injury or death in conflict scenarios. However, the article does not report any actual harm or incident caused by these AI systems yet; it focuses on the development and intended use of these systems as a deterrent and defense measure. Given the credible risk of harm from autonomous lethal systems and AI-guided military operations, this event is best classified as an AI Hazard rather than an AI Incident. It is not Complementary Information because it is not a response or update to a prior incident but a description of a new AI-enabled military capability with potential for harm.[AI generated]

The event involves an AI system (Meta's Muse Image) that uses user photos from public Instagram accounts to generate images. The automatic opt-in without explicit user consent and lack of notification about usage constitutes a violation of users' rights and privacy, which is a breach of obligations under applicable law protecting fundamental rights. This meets the criteria for an AI Incident under the definition of violations of human rights or breach of legal obligations. Although no direct physical harm is reported, the infringement on privacy and consent is a significant harm. Hence, the event is classified as an AI Incident.[AI generated]

The event involves the use of AI products in a fraudulent scheme that caused direct financial harm to victims, fulfilling the criteria for an AI Incident. The AI system's use was part of the deception leading to harm, and the harm is realized (fraud, financial loss). The involvement of AI is explicit and central to the scam's modus operandi. Hence, it is not merely a hazard or complementary information but an AI Incident.[AI generated]

The presence of AI systems is explicit, as Waymo's vehicles use advanced AI for autonomous driving and obstacle detection. The article details specific malfunctions and incidents where the AI systems failed to correctly interpret road conditions, leading to safety hazards and operational disruptions. These incidents have caused or could cause harm to people (passengers and others on the road) and have prompted regulatory action, fitting the definition of an AI Incident. The article also discusses ongoing efforts to validate and improve safety, but the primary focus is on realized harms and operational challenges, not just potential risks or responses.[AI generated]

The article explicitly mentions AI coding assistants, which are AI systems designed to assist in coding tasks. The flaw allows malicious repositories to run code on the developer's machine without proper authorization, which constitutes a direct harm to property and security. Since the AI systems' malfunction (the symlink flaw) directly leads to this harm, this qualifies as an AI Incident under the definition of harm to property caused by AI system malfunction.[AI generated]

The event involves AI systems (chatbots) whose use has directly led to harm in the form of misinformation and biased information dissemination, which affects public opinion and informed decision-making (harm to communities). The study documents actual occurrences of biased and disinformation sources being cited by AI chatbots, not just potential risks. The lack of transparency and the 'blackbox' nature of these systems exacerbate the harm. Therefore, this is an AI Incident as the AI systems' outputs have directly caused harm through misinformation and political bias.[AI generated]

The event explicitly involves AI meteorological models used by bloggers to generate and publish weather forecasts without authorization, violating the Meteorological Law of China. The AI system's outputs are used as public weather forecasts, which legally must be issued only by official meteorological agencies. This unauthorized use can cause misinformation and social harm, fulfilling the criteria for an AI Incident under violations of law and harm to communities. The involvement of AI in generating the forecasts and the resulting legal and social harm justifies classification as an AI Incident.[AI generated]

The event involves AI systems in autonomous vehicles and addresses safety concerns related to sensor configurations. However, no actual harm or incident has occurred; the bill is a preventive regulatory measure to mitigate plausible future risks. Therefore, this qualifies as an AI Hazard because the development and use of camera-only autonomous vehicles could plausibly lead to safety incidents, but no incident has yet materialized. The article does not describe a realized AI Incident, nor is it primarily about a societal or governance response without harm (which would be Complementary Information). It is not unrelated or a beneficial use event.[AI generated]

The AI system was used maliciously to generate a fake image to mislead authorities about the presence of a smoke detector, which is a safety device critical to protecting health and life. This misuse of AI directly contributed to a breach of legal obligations and posed a risk to tenant safety, fulfilling the criteria for an AI Incident. The harm is indirect but significant, as the AI-generated fake image was instrumental in the landlord's failure to comply with safety regulations, which could have led to injury or harm to tenants. Hence, the event is classified as an AI Incident.[AI generated]

The article explicitly identifies AI systems (LLM-based coding assistants and agents) as central to the attack vector. The attack exploits the AI systems' hallucination behavior to inject malicious commands that lead to large-scale infections and botnet formation, which can cause harm such as DDoS attacks and ransomware. These harms fall under harm to property, communities, and potentially critical infrastructure disruption. The involvement of AI is direct and pivotal, as the attack depends on the AI systems' flawed behavior. Hence, this is an AI Incident rather than a hazard or complementary information, as the harm is realized and the AI system's malfunction and use have directly led to it.[AI generated]

The article explicitly states that AI is being used to replace human roles, leading to the elimination of up to 1,800 jobs. This is a direct impact of AI use on employment, which is a significant harm to individuals and communities. The AI system's deployment is the cause of the job cuts, fulfilling the criteria for an AI Incident due to harm caused by AI use. There is no indication that the AI is solely a beneficial countermeasure or that the event is merely a complementary information update. Hence, the classification is AI Incident.[AI generated]