AIM: AI Incidents and Hazards Monitor

An AI system was explicitly used to create fabricated images that impersonated a police officer, which is a misuse of AI technology leading to harm in terms of deception and potential violation of rights and public trust. The AI-generated images directly contributed to the incident, and the legal authorities have intervened. This fits the definition of an AI Incident because the AI system's use directly led to a violation of law and harm to community trust, fulfilling criteria (c) violations of human rights or breach of obligations under applicable law and (e) other significant harms where AI's role is pivotal.[AI generated]

The article explicitly mentions the use of an unmanned submarine drone (an AI system) firing a lethal torpedo. While no harm has yet occurred, the system's lethal capability and military use imply a credible risk of future harm (injury, death, disruption). The event is not a realized incident but a demonstration of a potentially hazardous AI-enabled weapon system. Hence, it fits the definition of an AI Hazard rather than an AI Incident or other categories.[AI generated]

The lawsuit claims that OpenAI's ChatGPT logs could have been used to prevent a mass shooting, implying that the AI system's data was not utilized to avert harm. This connects the AI system's use (or failure to act on its outputs) to a significant harm (loss of life). Therefore, this event qualifies as an AI Incident because the AI system's involvement is directly linked to a serious injury/harm outcome.[AI generated]

The use of AI to generate fake voice and face content for fraudulent advertising directly leads to harm by deceiving consumers into purchasing counterfeit products, which constitutes realized harm. The AI system's malicious use is central to the incident, fulfilling the criteria for an AI Incident due to violations of consumer rights and harm to communities through fraud.[AI generated]

The article explicitly discusses the plausible future risks of advanced AI systems becoming too complex for humans to understand and control, which could lead to significant harms. It does not report any current harm or incident caused by AI, nor does it focus on responses or updates to past incidents. The presence of AI systems is clear, and the potential for harm is credible given the described capabilities and risks. Thus, the event is best classified as an AI Hazard, reflecting a credible risk of future harm from AI development and deployment.[AI generated]

The article focuses on potential risks and governance challenges related to the adoption of agentic AI systems, emphasizing plausible future harms if inadequate controls are applied. It does not report any actual harm or incident resulting from AI use, nor does it describe a specific event where AI caused injury, rights violations, or other harms. Therefore, it fits the definition of an AI Hazard, as it outlines circumstances where AI use could plausibly lead to harm, but no harm has yet occurred. It is not Complementary Information because it is not updating or responding to a past incident but rather providing a forward-looking risk assessment.[AI generated]

The event involves an AI system (Waymo's autonomous vehicle) operating in a complex environment (farmers market during setup). While no injury or damage occurred, the vehicle's presence in a restricted area with people around presents a credible risk of harm if the AI system were to malfunction or misinterpret the environment. This fits the definition of an AI Hazard, as the event plausibly could lead to an AI Incident in the future. Since no actual harm occurred, it is not an AI Incident. The event is not merely complementary information or unrelated, as it involves a specific AI system's operation with potential safety implications. It is also not a beneficial use, as the AI system is not deployed solely as a countermeasure to an external harm.[AI generated]

The article centers on a leaked EU proposal that would allow AI companies to use personal data extensively without consent, potentially breaching data protection rights. While no direct harm has yet occurred, the proposal's adoption could plausibly lead to violations of fundamental rights protected by law, fitting the definition of an AI Hazard. The AI systems' data processing is central to the risk, but since the event is about a legislative proposal and not an actual incident, it is not an AI Incident. It is not Complementary Information because it is not an update or response to a past incident but a new potential risk. It is not Beneficial Use or Unrelated.[AI generated]

The article explicitly mentions the use of AI-generated videos by both BJP and SP to defame political opponents, which is a direct use of AI systems causing harm to individuals and communities through misinformation and reputational damage. The harm is realized as the videos are widely circulated and have prompted official responses and warnings. This fits the definition of an AI Incident because the AI system's use has directly led to harm to communities and violations of rights related to information and reputation. There is no indication that this is merely a potential risk or a response to an external harm, so it is not an AI Hazard or Beneficial Use. It is also not Complementary Information since the main focus is on the harmful AI-generated content itself.[AI generated]

The event involves the use of AI systems to analyze and combine voice samples collected via silent calls to create convincing voice clones for fraudulent purposes. This use of AI directly contributes to financial fraud, which is a harm to persons and their property. The article reports that such frauds have increased and are actively occurring, fulfilling the criteria for an AI Incident. The AI system's role is pivotal in enabling the voice cloning and impersonation that facilitate the fraud.[AI generated]

The McKinsey report explicitly states that 30% of companies experienced productivity decline after adopting agentic AI tools, indicating realized harm caused by AI use. The AI system is agentic AI for autonomous coding, clearly an AI system. The harm is operational and economic, which qualifies as harm to organizations and communities under the framework. The event is not merely a potential risk or a complementary update but a documented negative outcome directly linked to AI use. Therefore, it qualifies as an AI Incident.[AI generated]

The creation and dissemination of deepfake sexual videos using AI technology constitutes a violation of human rights and privacy, which is a breach of applicable law protecting fundamental rights. The harm is realized, as victims have reported impact and legal action is underway. The AI system's role is pivotal as it enabled the generation of realistic fake videos causing reputational and emotional harm. Therefore, this event qualifies as an AI Incident.[AI generated]

The article explicitly discusses AI systems being used maliciously or misconfigured, leading to actual cybersecurity incidents such as data breaches, unauthorized access, and fraud. The presence of AI-enabled threats like malicious AI plugins and AI bots that manipulate social processes confirms direct harm caused by AI systems. The lack of governance policies enabling 'shadow AI' usage further contributes to these harms. Since these harms are realized and directly linked to AI system use and misuse, the event qualifies as an AI Incident under the framework definitions.[AI generated]

The event involves the use of AI systems (deepfake and AI-generated content) in the malicious use context to perpetrate phishing scams and fraud, which directly harms individuals by stealing personal and financial information. This fits the definition of an AI Incident because the AI system's use has directly led to harm to people (financial harm and violation of privacy). The article warns about ongoing scams, indicating harm is occurring, not just a potential risk. Therefore, the classification is AI Incident.[AI generated]

The event involves AI systems (GPT-6 Astra and others) controlling robots and executing dangerous commands, which directly relates to the AI system's use and malfunction. The experiments showed that the AI often failed to filter out harmful instructions, leading to robot actions that could cause injury or property damage. Even though no actual harm occurred during the tests, the AI's behavior demonstrates a direct link to potential physical harm, fulfilling the criteria for an AI Incident. The article focuses on the AI's role in causing or enabling these dangerous actions, not just potential future harm or general AI research, so it is not an AI Hazard or Complementary Information. It is not Beneficial Use since the AI is not preventing harm but potentially causing it, and it is not Unrelated as the AI system's involvement is explicit and central.[AI generated]

The Tesla FSD is an AI system providing driver assistance but not full autonomy. The driver was asleep, relying on the system beyond its intended use, which is a misuse of the AI system. The police intervention prevented potential harm. Since no injury or accident occurred, it is not an AI Incident. However, the event demonstrates a plausible risk of harm due to misuse of AI-assisted driving, fitting the definition of an AI Hazard. The event is not merely general news or complementary information because it involves a specific event with plausible future harm due to AI misuse.[AI generated]

The event involves AI in an economic context, with concerns about AI potentially deepening supply-demand imbalances in China. However, no actual harm or incident has occurred; the article reflects a forecast or warning about possible future economic effects. This fits the definition of an AI Hazard, as it plausibly could lead to economic harm but no direct or indirect harm has yet materialized. The article does not focus on a specific AI Incident, Complementary Information, Beneficial Use, or unrelated news.[AI generated]

The article describes the construction of a large unmanned submarine mothership designed to deploy AI-enabled unmanned underwater vehicles. These vehicles inherently involve AI systems for autonomous operation. The event is about the development and potential use of such AI systems in military applications, which could plausibly lead to harms such as conflict escalation or damage to critical infrastructure. No actual harm or incident is reported, so it is not an AI Incident. The event is not a societal or governance response, nor is it a beneficial use or unrelated news. Hence, it fits the definition of an AI Hazard due to the credible risk posed by the deployment of AI-enabled unmanned military vessels.[AI generated]

The event explicitly involves AI systems generating deepfake videos used in political campaigns, which have directly led to harm by misrepresenting a candidate's statements and potentially misleading voters. This constitutes harm to communities and a violation of rights under the definitions. The complaint and police report indicate the harm is realized, not just potential. Therefore, this is an AI Incident rather than a hazard or complementary information.[AI generated]

The article clearly involves AI systems (chatbots like ChatGPT, Gemini, Claude) used for financial advice. The study shows these AI systems often give wrong answers, which can mislead users and cause financial harm. This meets the definition of an AI Incident because the AI's use has directly or indirectly led to harm (financial loss) to people. The harm is realized, not just potential, as the article cites specific examples of incorrect advice that could cause severe financial outcomes. Therefore, this is an AI Incident.[AI generated]