AI Outperforms Virologists, Raising Bioweapon Concerns

The article explicitly involves AI systems (ChatGPT, Claude, OpenAI's o3, Google's Gemini 2.5 Pro) demonstrating advanced capabilities in virology-related problem-solving. While no direct harm has occurred yet, the potential misuse of these AI models by non-experts to create deadly bioweapons represents a credible and plausible risk of harm to health and safety. Therefore, this event fits the definition of an AI Hazard, as it could plausibly lead to an AI Incident involving injury or harm to people.

The article explicitly mentions AI systems (ChatGPT, Gemini) and their advanced problem-solving capabilities in virology, which could be misused by malicious individuals to develop bioweapons. Although no incident of harm has occurred, the potential for such harm is credible and significant. This fits the definition of an AI Hazard, as the development and use of these AI systems could plausibly lead to an AI Incident involving harm to health and safety. Therefore, the event is best classified as an AI Hazard rather than an Incident or Complementary Information.

The event involves AI systems (large language models like OpenAI's o3 and Google's Gemini) demonstrating expert-level capabilities in virology lab procedures, which is explicitly stated. The study highlights dual-use concerns where these AI capabilities could plausibly lead to serious harm (creation of bioweapons), fulfilling the definition of an AI Hazard. Although no actual harm has been reported yet, the credible risk of misuse and calls for regulatory action confirm the plausible future harm. The article also discusses mitigation efforts by AI developers, but the primary focus is on the potential for harm rather than realized incidents. Therefore, this event is best classified as an AI Hazard.

The event involves AI systems (large language models like ChatGPT, Claude, OpenAI's o3) whose use in virology lab problem-solving has directly led to concerns about the potential for harm through bioweapon creation. The AI's role is pivotal as it provides practical expertise that was previously inaccessible to non-experts, thus enabling a credible risk of serious harm to public health and safety. Although no specific incident of bioweapon creation is reported, the article highlights realized capabilities that have already changed the risk landscape and the direct involvement of AI in enabling this risk. This fits the definition of an AI Hazard with a strong potential for harm, but given the direct link to realized AI capabilities and the discussion of mitigation and regulatory responses, it is best classified as an AI Hazard rather than an incident since no actual harm event has yet occurred.

The event involves AI systems (large language models) whose development and use have led to a credible risk of harm through potential bioweapon creation, which fits the definition of an AI Hazard. No actual harm has yet occurred, but the plausible future harm is clearly articulated and linked to the AI's capabilities. The article also discusses mitigation measures, but the primary focus is on the potential for harm rather than realized incidents or governance responses alone.

The event involves AI systems (large language models) whose use has directly led to a significant risk of harm by enabling potentially dangerous bioweapon creation. Although no actual bioweapon creation incident is reported, the demonstrated AI capabilities and the expressed concerns about misuse constitute a credible and serious risk of harm to public health and safety. This fits the definition of an AI Hazard because the AI's development and use could plausibly lead to an AI Incident involving injury or harm to groups of people. The article also discusses responses and mitigation efforts, but the primary focus is on the risk posed by the AI capabilities themselves, not on a realized harm or an incident. Therefore, the classification is AI Hazard.