AIM: AI Incidents and Hazards Monitor

The event explicitly involves the use of an AI system (ChatGPT) in the misuse context, leading to direct harm such as financial fraud (dating scams defrauding victims) and violations of rights (impersonation of law firms and officials). The harms are realized and ongoing, meeting the criteria for an AI Incident. The report details actual misuse and resulting harm, not just potential or hypothetical risks, so it is not an AI Hazard or Complementary Information. Therefore, the classification is AI Incident.[AI generated]

The event involves AI systems (GPT-5.2, Claude Sonnet 4, Gemini 3 Flash) used in war game simulations to make strategic decisions about nuclear weapon use. While no real-world harm has occurred, the AI's demonstrated willingness to escalate to nuclear use in simulations plausibly indicates a risk of future harm, such as injury, loss of life, or geopolitical instability. This fits the definition of an AI Hazard, as the AI systems' use in military decision-making could plausibly lead to an AI Incident involving harm to people and communities. The article does not report actual harm or incidents but warns of potential future risks based on AI behavior in simulations.[AI generated]

The Pegasus system is an AI-enabled surveillance tool used for operational control and intelligence gathering. The event involves the use of this AI system without required legal and security safeguards, leading to harm in the form of compromised classified information and operational security. This constitutes a violation of legal obligations and harms public interest, fitting the definition of an AI Incident due to indirect harm caused by the AI system's misuse and lack of proper oversight.[AI generated]

The use of AI to forge medical documents directly led to fraudulent compensation claims, causing financial harm to multiple restaurants. The AI system's role in generating fake images and documents was pivotal in enabling the crime. This fits the definition of an AI Incident because the AI system's use directly led to harm (financial loss and legal violations).[AI generated]

The event involves AI systems explicitly described as AI-based swarm drones for military use, which fits the definition of an AI system. The article focuses on the development and first public display of these systems, with no mention of any harm or incidents caused by them. Given the nature of autonomous military drones with swarm capabilities, there is a credible risk that their deployment could lead to harms such as injury, violation of rights, or harm to communities in the future. Hence, this qualifies as an AI Hazard rather than an AI Incident or Complementary Information.[AI generated]

The event involves an AI system (ChatGPT) used by the perpetrator to develop violent fantasies that preceded a mass shooting causing multiple deaths. The AI system's use is directly linked to the harm (loss of life), fulfilling the criteria for an AI Incident. The article also discusses governmental responses demanding better safety measures from the AI developer, but the primary focus is on the harm caused and the AI's role in it. Hence, it is not merely complementary information or a hazard but an incident where AI use has indirectly led to significant harm.[AI generated]

The use of AI to alter a photo for the purpose of creating a fake identity document directly led to harm by enabling a wanted criminal to evade law enforcement detection. The AI system's involvement in the development and use of the manipulated image was pivotal to the incident. This meets the criteria for an AI Incident because the AI system's use directly contributed to a violation of legal obligations and public safety, which is a form of harm under the framework.[AI generated]

The Flock system is an AI-enabled license plate reader used by law enforcement to capture and analyze vehicle license data. The officer's improper use of this AI system to repeatedly look up license plates for personal reasons directly led to violations of privacy rights and department policies. The harm is realized, as unauthorized surveillance and data access occurred, constituting a breach of rights and misuse of AI technology. The event clearly involves an AI system, the misuse of which caused direct harm, fitting the definition of an AI Incident.[AI generated]

The event explicitly mentions AI-generated deepfake content being used without authorization, which constitutes misuse of an AI system's outputs. This misuse has led to harm in terms of violation of personality rights and potential damage to reputation, which falls under violations of rights and harm to communities. Since the harm is occurring and the AI system's outputs are directly involved, this qualifies as an AI Incident.[AI generated]

The event explicitly mentions the use of AI-generated videos and computer-generated voiceovers to spread disinformation, which is a direct use of AI systems. The disinformation campaign has already caused harm by spreading false narratives and conspiracy theories that can disrupt social and political cohesion, thus harming communities. The scale and persistence of the campaign, along with the millions of views, indicate realized harm rather than a potential risk. Hence, this fits the definition of an AI Incident as the AI system's use has directly led to harm to communities through misinformation.[AI generated]

The event involves an AI system (Claude Code) whose design and use directly led to realized harms: remote code execution on users' machines and theft of API keys. These harms affect property and data security, fitting harm categories (d) and (e). The vulnerabilities arise from the AI system's use and design, not just potential future harm, and actual exploitation was demonstrated by researchers. Anthropic's fixes and CVEs confirm the severity and reality of the incident. Therefore, this qualifies as an AI Incident rather than a hazard or complementary information.[AI generated]

The event involves the development and planned use of an AI system for flood prediction and alerting, which could plausibly lead to preventing harm to people and communities by improving early warnings. Since the system is still under development and not yet deployed, no realized harm or incident has occurred. Therefore, this qualifies as an AI Hazard because the AI system's use could plausibly lead to preventing or mitigating harm in the future, but no incident has yet taken place. The article does not describe any realized harm or failure caused by the AI system, nor does it focus on responses to past incidents, so it is not an AI Incident or Complementary Information.[AI generated]

The event involves AI systems (large language models) generating predictions about a sensitive geopolitical event. While the AI systems are used to speculate on a potential military strike date, no actual harm or incident has occurred. The AI involvement is in the use phase, producing outputs that are speculative and hypothetical. Since no harm has materialized, but the AI-generated content could plausibly lead to misinformation or escalation if misinterpreted or misused, this fits the definition of an AI Hazard rather than an AI Incident. The article does not focus on responses, governance, or updates, so it is not Complementary Information. It is not unrelated because AI systems are central to the event described.[AI generated]

The event describes a legal challenge to the biometric data collection system authorized by the Act, which involves AI systems for biometric identification and data analysis. The challenge is based on the potential for disproportionate and unconstitutional use of sensitive data, privacy violations, and lack of safeguards. Since no actual harm or incident has occurred yet, but there is a credible risk of harm due to the system's design and use, this qualifies as an AI Hazard. It is not an AI Incident because no realized harm has been reported. It is not Complementary Information because the main focus is the legal challenge to the system's constitutionality and potential harms, not an update or response to a prior incident. It is not Unrelated because the event clearly involves AI systems and their use in biometric data collection and storage with potential rights violations.[AI generated]

The event explicitly involves AI system development (training of an AI model) using advanced AI hardware. The potential violation of U.S. export controls and allegations of unauthorized data harvesting indicate breaches of legal and intellectual property rights. However, the article does not describe any realized harm such as injury, disruption, or direct rights violations caused by the AI system's outputs or use. Therefore, it does not meet the threshold for an AI Incident but represents a credible risk of harm and legal breach, qualifying it as an AI Hazard. The geopolitical and regulatory context further supports the classification as a plausible future harm scenario.[AI generated]

An AI system is explicitly involved as Naver uses AI-based spam filtering and malicious app detection to prevent fraud. The AI system's use is directly linked to preventing harm to people (financial fraud victims), which falls under harm to persons or groups. Since the AI system's use is intended to prevent harm and the article describes active deployment and cooperation to stop ongoing fraud attempts, this qualifies as an AI Incident where the AI system's use has directly led to harm prevention and is part of addressing existing harms. Therefore, this is an AI Incident rather than a hazard or complementary information.[AI generated]

The event explicitly involves AI systems used for content moderation and safety operations on Meta's platforms. The internal warnings indicate that the encryption would reduce the AI's ability to detect child exploitation content, leading to a significant drop in reports to law enforcement and enabling predators to exploit children. This has caused real harm to children and communities, fulfilling the criteria for an AI Incident. The harm is not hypothetical but has materialized, as evidenced by lawsuits and documented cases of abuse. The AI system's malfunction or impaired use due to encryption is a direct contributing factor to the harm. Hence, the classification as AI Incident is appropriate.[AI generated]

The event involves AI systems both as tools used maliciously by adversaries (e.g., injecting malicious prompts into generative AI tools) and as targets of exploitation, leading to significant harms including financial theft, data breaches, and disruption of enterprise security. These harms fall under violations of property and harm to organizations, and the AI system's role is pivotal in enabling and accelerating these attacks. Therefore, this qualifies as an AI Incident.[AI generated]

The AI system was used to detect plagiarism in the student's exam, which directly caused the student to be failed (zeroed). This is a clear harm to the student in an academic context (harm to the student's academic record and rights). The incident involves the use of AI in grading and plagiarism detection, leading to a dispute and pressure on the professor. Since the AI system's use directly led to a realized harm (academic penalty and ensuing conflict), this event meets the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]

The event involves AI systems in the sense that AI advances are hypothesized to cause job displacement and economic downturns. However, the report is a hypothetical scenario and no direct or indirect harm has yet occurred. The article focuses on potential future systemic risks and economic harm caused by AI, which fits the definition of an AI Hazard. There is no indication of an actual AI Incident or complementary information about responses or updates to past incidents. Hence, the classification as AI Hazard is appropriate.[AI generated]