AIM: AI Incidents and Hazards Monitor

The article explicitly mentions the use of an AI system being trained with employee activity data collected via a new tracking tool. While employees express concerns about potential job cuts and privacy implications, no actual harm or rights violations have been documented as having occurred. The tracking for AI training purposes could plausibly lead to harms such as labor rights violations or privacy breaches, making this an AI Hazard rather than an AI Incident. It is not Complementary Information because the main focus is on the new tracking tool and its implications, not on responses or updates to prior incidents. It is not Unrelated because the event clearly involves AI system development and use with potential for harm.[AI generated]

The event involves the use of generative AI systems to create a fake influencer persona that deceived users and generated income through fraudulent means. The AI system's outputs were central to the deception and monetization, directly causing harm to the users who were misled and financially exploited. The account was removed for fraudulent activity, confirming the harm occurred. This fits the definition of an AI Incident as the AI system's use directly led to harm (financial and trust-related) to groups of people.[AI generated]

The event explicitly involves the use of AI to generate legal citations, which were fabricated ('hallucinations'), leading to errors in a court filing. This directly caused harm by misleading the court and opposing counsel, constituting a violation of legal and professional standards. The AI system's malfunction or misuse is central to the incident. The harm is realized, not just potential, as the false citations were submitted and discovered, prompting an apology and review. Hence, it meets the criteria for an AI Incident under violations of legal obligations and harm to the judicial process.[AI generated]

The article explicitly mentions the use of AI-based autonomous and semi-autonomous systems for military purposes, which qualifies as AI system involvement. The event concerns the development and planned deployment of these systems, not a realized harm. However, autonomous weapons and military AI systems inherently carry credible risks of causing injury, disruption, or other harms. Since no actual harm is reported yet, but the plausible future harm is clear, this fits the definition of an AI Hazard. It is not Complementary Information because it is not an update or response to a prior incident, nor is it unrelated as it directly involves AI systems with potential for harm.[AI generated]

The article explicitly mentions an AI system (Anthropic's Mythos) capable of identifying and exploiting software vulnerabilities. The Bundesbank chief warns about the potential for malicious use, which could lead to disruption of critical infrastructure (financial institutions). Since no actual incident has occurred yet but there is a credible risk of harm, this qualifies as an AI Hazard rather than an AI Incident or Complementary Information.[AI generated]

The event explicitly involves an AI system generating a deepfake video, which is a clear AI application. The deepfake has been used maliciously to impersonate Wendie Renard and promote a scam, directly causing reputational harm and posing financial risks to people targeted by the video. The harm is realized (identity theft, potential financial fraud), and the AI system's role is pivotal in enabling this harm. Therefore, this qualifies as an AI Incident under the framework.[AI generated]

The article explicitly mentions the development and intended use of AI systems (smart glasses with facial recognition and biometric databases) by DHS/ICE for surveillance purposes. The potential harms include violations of civil rights, privacy, and mass surveillance, which are serious human rights concerns. However, the article does not report any actual harm or incident resulting from the use of these glasses yet, only the plans and concerns about their future use. Thus, it fits the definition of an AI Hazard, where the AI system's use could plausibly lead to an AI Incident but has not yet done so.[AI generated]

The AI system is explicitly involved as it generates the singer's music and image. The harm arises from the AI-generated content reinforcing racist clichés and stereotypes about the Roma minority, which is a violation of human rights and harms the community. The event describes realized harm through social and cultural impacts, including criticism from Roma activists and musicians, and the perpetuation of latent racism. Hence, it meets the criteria for an AI Incident due to indirect harm caused by the AI system's outputs.[AI generated]

The article explicitly involves AI systems (AI code generation tools) whose outputs (AI-generated code) have directly led to bugs and defects requiring additional review and fixes, causing increased workload and quality concerns. These constitute realized harms related to software quality and reliability, which fall under harm to property and disruption of operations. The survey data confirms that these harms are occurring and significant. Hence, this qualifies as an AI Incident rather than a hazard or complementary information, as the harms are realized and directly linked to AI system use.[AI generated]

The event involves the use of AI systems (ChatGPT and LLMs) and their impact on human brain activity and cognitive skills. The study shows a correlation between AI reliance and diminished critical thinking, which is a form of potential harm to individuals' cognitive health and educational development. However, the article does not report any realized injury, rights violation, or other direct harm caused by the AI system's malfunction or misuse. The harm is potential and plausible, related to future educational and cognitive risks. Thus, it fits the definition of an AI Hazard rather than an AI Incident or Complementary Information. It is not unrelated because it clearly involves AI systems and their effects.[AI generated]

The event involves the use of AI systems to generate fake job offers and messages that trick victims into giving money or sensitive personal information, causing realized harm. The AI system's role is pivotal in making the scams credible and effective, directly leading to harm to individuals' finances and privacy. Therefore, this qualifies as an AI Incident under the definition of harm to persons and communities caused by AI misuse.[AI generated]

An AI system was explicitly used by the police to identify and locate the suspect vehicle and driver after a hit-and-run incident causing injury to a person. The AI system's involvement was in the use phase, aiding in the investigation and arrest. The harm (injury to the pedestrian) has occurred, and the AI system played a pivotal role in addressing the incident. Therefore, this qualifies as an AI Incident due to the direct link between the AI system's use and the harm event.[AI generated]

The involvement of AI is explicit through the mention of deepfakes, which are AI-generated synthetic media. The harm relates to violation of personality rights and intellectual property rights due to unauthorized commercial use of AI-generated content. Since the court order is a response to an ongoing harm (unauthorized use of AI to exploit the actor's persona), this constitutes an AI Incident involving violation of rights. The event is not merely a general AI-related update or a potential future risk but a concrete legal action addressing realized harm caused by AI misuse.[AI generated]

The event involves an AI system (Anthropic's Mythos) with advanced capabilities that could plausibly lead to harm by destabilizing banking systems through cybersecurity vulnerabilities. However, the article only describes regulatory monitoring and risk assessment without any realized harm or incident. Therefore, this constitutes an AI Hazard, as the AI system's development and potential use could plausibly lead to an AI Incident in the future, but no direct or indirect harm has occurred yet.[AI generated]

The article explicitly involves an AI system (Claude Mythos Preview) capable of discovering software vulnerabilities autonomously, which could be exploited to cause cyberattacks on critical financial infrastructure. Although no actual incident or harm has occurred yet, the article outlines credible scenarios where such AI capabilities could lead to significant harm, including disruption of financial services and loss of trust, which fall under harm categories (b) and (d). The focus is on plausible future harm and preparedness rather than a realized incident, fitting the definition of an AI Hazard. The article also discusses governance and mitigation strategies but the primary subject is the potential risk posed by the AI system, not just complementary information about responses. Hence, the classification is AI Hazard.[AI generated]

The autonomous driving system is an AI system as it uses AI to perceive and make driving decisions. The event involves the use and malfunction of this AI system, which directly led to physical harm and fatalities, fulfilling the criteria for an AI Incident. The concealment of these incidents and the court ruling demonstrate that harm has materialized, not just potential harm. Therefore, this event is classified as an AI Incident.[AI generated]

An AI system is involved as Lovable is an AI app-building platform handling AI chat histories and code projects. The event stems from the use and design of the AI system's visibility settings, which led to unauthorized access to sensitive data, including AI chat histories and customer records. This constitutes a violation of privacy and potentially breaches obligations to protect user data, which falls under harm to rights and possibly harm to individuals. Although the company denies a breach, the exposure of sensitive data due to unclear documentation and design is a realized harm. Therefore, this qualifies as an AI Incident because the AI system's use and design directly led to harm through unauthorized data exposure.[AI generated]

The article explicitly describes the use of AI systems (AI agents orchestrated via MCP) in a confirmed cyber-espionage campaign that targeted high-value organizations, causing harm through unauthorized data access and exploitation. The MCP flaw is a systemic architectural vulnerability in AI system integration, directly enabling these attacks. The harm is realized and significant, involving breaches of security and potential violations of rights and property. The involvement of AI is central and pivotal to the incident, as the AI agents autonomously conducted the intrusion lifecycle. This meets the criteria for an AI Incident because the AI system's use and the architectural flaw directly led to harm. The article also discusses broader systemic risks and governance responses but the primary focus is on the realized harm from AI misuse.[AI generated]

The article explicitly mentions the use of generative AI tools by scammers to impersonate tax officials, making scams more efficient and prevalent. This involves the use of AI systems in the malicious use category, directly leading to harm through tax fraud and identity theft. The harm includes theft of sensitive information and financial loss, which fits the definition of harm to persons and violation of rights. Hence, this is an AI Incident due to the realized harm caused by AI-enabled fraudulent activities.[AI generated]

The article explicitly mentions advanced AI systems with capabilities to find and exploit cybersecurity weaknesses, which could plausibly lead to harm such as disruption of critical financial infrastructure or harm to financial security. However, the article only reports warnings and concerns about potential risks, not an actual realized incident. Therefore, this qualifies as an AI Hazard, reflecting a credible future risk stemming from AI system capabilities and their potential misuse or exploitation.[AI generated]