AIM: AI Incidents and Hazards Monitor

The article involves AI systems in the context of economic and labor market impacts but does not report a realized harm or incident caused by AI. The concerns raised are about plausible future harms such as job displacement and inequality, which aligns with the definition of an AI Hazard. There is no mention of a specific AI malfunction or misuse causing harm, nor is the article focused on responses or updates to past incidents, so it is not Complementary Information. Therefore, the event is best classified as an AI Hazard due to the credible risk of AI exacerbating economic disparities and job losses in the future.[AI generated]

The AI system's involvement is in the discovery of vulnerabilities, which is a use of AI in cybersecurity. No direct or indirect harm has yet occurred as a result of the AI system or the vulnerabilities; rather, the article highlights the response to a potential threat. Therefore, this event represents a plausible future risk (hazard) due to the vulnerabilities discovered, but since no harm has materialized, it is not an incident. The focus on mitigation and patching indicates a hazard scenario where harm could plausibly occur if not addressed.[AI generated]

The event involves the use and damage of AI systems' infrastructure (data centres hosting AI supercomputers and services) which has directly led to widespread disruption of AI services and setbacks in AI development. The harm includes disruption of critical infrastructure (cloud services, banking, e-commerce), harm to AI development capabilities, and consequential economic and societal impacts. The AI system involvement is explicit (Yandex's AI services and supercomputers), and the harm is realized, not just potential. Thus, it meets the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]

The event involves the use of AI systems (e.g., AI tools for generating and spreading disinformation and deepfake videos) in an active disinformation campaign that has already caused harm by undermining trust in political processes and damaging reputations. The harm to communities and political rights is direct and ongoing. The article explicitly mentions AI's role in accelerating and refining the spread of manipulative content, fulfilling the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]

The article explicitly states that Barclays is using more AI, which has made some work processes more efficient, leading to a 5% reduction in its workforce. This is a direct use of AI causing harm in the form of job losses, which is a significant social and economic harm to individuals and communities. Therefore, this qualifies as an AI Incident under the definition of harm to communities or groups of people resulting from AI use.[AI generated]

The article clearly involves AI systems used for game development, which is accelerating the creation of crypto games. The potential harms include financial exploitation and speculative losses for players, which are plausible future harms. However, no actual harm or incident is reported as having occurred yet. The discussion is largely predictive and speculative, focusing on the risks of a flood of low-quality games enabled by AI. This fits the definition of an AI Hazard, where AI use could plausibly lead to harm but has not yet directly or indirectly caused harm. It is not Complementary Information because the article is not about responses or updates to a past incident, nor is it Beneficial Use or Unrelated.[AI generated]

The event involves the use of an AI system ('Artex') in the malicious cyberattack campaign that caused actual data breaches and unauthorized access to sensitive information, including in South Korean financial institutions. This constitutes a violation of rights (privacy and data protection) and harm to communities and property (data and organizational security). The AI system's use was pivotal in automating and enabling the attack, thus directly leading to realized harm. Therefore, this qualifies as an AI Incident under the OECD framework.[AI generated]

The robot described is clearly an AI system, given its humanoid autonomous capabilities, physical interaction, and advanced processing (e.g., Nvidia Jetson AGX Thor). The event involves the development and demonstration of this AI system. No actual harm or incident is reported, so it is not an AI Incident. However, the article discusses the robot's potential use as a bodyguard and its impressive physical power, which could plausibly lead to harm if misused or malfunctioning. This fits the definition of an AI Hazard, as the development and potential deployment of such robots could plausibly lead to injury, property damage, or other harms in the future. The article does not focus on societal or governance responses or updates, so it is not Complementary Information. It is not Beneficial Use because the robot is not currently deployed solely as a countermeasure to external harm without introducing new harm. Therefore, the correct classification is AI Hazard.[AI generated]

The AI system (smart glasses with AI functionality) was explicitly used to commit an act of cheating during a national exam, which is a violation of legal obligations and disrupts the exam's management and operation. This constitutes a direct harm under the framework's category (c) violations of human rights or breach of obligations under applicable law, and (b) disruption of critical infrastructure management (exam administration). Therefore, this event qualifies as an AI Incident because the AI system's use directly led to realized harm (legal violation and disruption).[AI generated]

The article does not describe a specific AI system causing direct or indirect harm, nor does it report an actual incident or malfunction. Instead, it presents a forecast and analysis of potential future labor market disruptions due to AI, which constitutes a plausible risk but not a realized harm. Therefore, this qualifies as an AI Hazard because it highlights a credible potential for AI to cause significant labor displacement and associated harms in the future, but no actual harm has yet occurred as per the article.[AI generated]

The use of AI to create manipulated videos featuring political figures with false statements directly leads to harm to communities by spreading misinformation and potentially violating rights related to reputation and truthful political discourse. The AI system's use in generating this content is central to the harm. Therefore, this qualifies as an AI Incident due to realized harm from AI-generated misinformation and reputational damage in a political context.[AI generated]

The social media platforms operated by Meta use AI systems to design and deliver content feeds that can create addictive behaviors. The lawsuit alleges that this AI-driven design has directly led to harm (addiction and exposure to harmful content) in children and adolescents, which fits the definition of an AI Incident involving harm to health and well-being of groups of people. Therefore, this event qualifies as an AI Incident.[AI generated]

The article involves AI systems (AI agents) that have been used to access bank accounts and perform transactions, sometimes independently of user instructions. This use of AI has led to concerns about unauthorized actions that could harm customers financially or compromise banking operations. Although no specific harm incidents are detailed, the banks' blocking of these AI agents indicates a response to realized or imminent risks. Since the article reports on actual unauthorized AI agent activity and the banks' response to prevent harm, this qualifies as an AI Incident due to the direct or indirect harm (or risk thereof) caused by AI agents acting without proper authorization.[AI generated]

The article explicitly states that AI systems were used to discover a critical security vulnerability in PHP, which is foundational to about 70% of websites globally. The vulnerability's discovery led to widespread temporary outages as sites updated their software to prevent exploitation. This disruption constitutes harm to property and services (websites going offline), and the potential for hacking represents a direct threat to security and privacy. The AI's involvement in detecting the vulnerability is central to the event, and the resulting harm (service disruption and cybersecurity risk) is realized, not merely potential. Hence, this is an AI Incident rather than a hazard or complementary information.[AI generated]

The AI system in question is Flock Safety's automated license plate readers and AI surveillance cameras, which clearly fit the definition of AI systems due to their automated data capture and analysis capabilities. The event documents the use of these AI systems for mass surveillance, which has led to privacy concerns and potential violations of human rights. The mapping tool reveals the extent of this surveillance, indicating that the AI system's deployment has already caused harm by enabling widespread, unchecked tracking of individuals. This meets the criteria for an AI Incident because the AI system's use has directly led to harm related to violations of rights and harm to communities. The event is not merely a hazard or complementary information, as the harm is ongoing and realized.[AI generated]

The article explicitly mentions the use of an AI system (ARTEX) in the malicious use context to conduct penetration testing attacks that led to a supply chain compromise and breach of Microsoft 365 tenants. This breach constitutes harm to property and possibly to communities due to data exposure. The AI system's involvement is direct and pivotal in enabling the attack at scale and speed. Therefore, this qualifies as an AI Incident.[AI generated]

The compromised LiteLLM package is an AI system interface that was maliciously altered to include malware stealing sensitive credentials related to AI and cloud services. The attack directly led to harm by exposing secrets and enabling remote code execution, which fits the definition of an AI Incident due to violations of rights and harm to property and communities. The presence of the AI system is explicit, and the harm is realized, not just potential. Therefore, this event is classified as an AI Incident.[AI generated]

The article explicitly involves AI systems (advanced AI models) used to generate cyberattack plans against critical infrastructure, which is a clear AI system involvement. The event stems from the use of AI systems in a malicious context (planning attacks). While no actual harm has occurred yet, the research demonstrates that these AI systems could plausibly lead to disruption of critical infrastructure, meeting the criteria for an AI Hazard. Therefore, this event is best classified as an AI Hazard rather than an AI Incident, as the harm is potential, not realized.[AI generated]

The event involves a vehicle equipped with AI-enabled intelligent driving and emergency call systems. The fatal accident and fire caused the death of the driver, constituting harm to a person. The AI system's involvement is indirect but significant, as the intelligent driving system's safety and the vehicle's AI features are questioned in the aftermath. The incident is not merely a product announcement or general news; it reports a realized harm linked to an AI system's use or malfunction (or at least its safety implications). The controversy over the company's explanation and data transparency further highlights the AI system's role in the incident. Hence, the event meets the criteria for an AI Incident.[AI generated]

Meta's content moderation relies on AI algorithms that automatically remove or block content. The arbitrary removal of at least 31 journalistic posts, which were not propaganda or in violation of electoral rules, indicates a malfunction or misuse of the AI system leading to censorship. This has caused harm by restricting access to information during a critical democratic process, constituting a violation of rights and harm to communities. The involvement of AI in the content removal and the resulting harm meet the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]