AI-Generated Fake Citations Undermine Scientific Research Integrity

The article explicitly identifies AI systems (LLMs) generating false citations that have been found in a large number of scientific papers, directly undermining the reliability of scientific research. This constitutes harm to communities (the scientific community and society relying on research) and breaches academic integrity, which aligns with violations of intellectual property and fundamental rights. The harm is realized and widespread, not merely potential. The involvement of AI in generating these fake citations is direct and causal. Hence, the event meets the criteria for an AI Incident rather than a hazard or complementary information.

The article involves the use of AI systems (e.g., AI analyzing crop quality and growth via imaging technologies) in the context of smart agriculture. However, it describes a research and development effort aimed at preventing future harm from climate change rather than an event where AI use has directly or indirectly caused harm. The potential harms (crop failure due to climate change) are future risks being addressed proactively. Therefore, this event is best classified as Complementary Information, as it provides context on AI's role in addressing environmental challenges and does not report an AI Incident or AI Hazard.

The article discusses the development and open-source release of a software platform that supports autonomous vehicles by integrating roadside sensor data. While this involves AI systems (autonomous driving and sensor data integration), the event itself is about the software release and research collaboration, without any mention of harm or incidents caused by the AI system, nor any plausible future harm directly indicated. Therefore, it does not describe an AI Incident or AI Hazard. It is not a routine product launch but a research software release, which provides complementary information about AI development and ecosystem progress.

The event explicitly involves AI systems (LLMs) generating hallucinated citations that have been published in scientific papers, which is a direct use of AI leading to harm. The harm includes violations of intellectual property rights (fabricated citations infringe on proper attribution and scholarly norms) and harm to communities (scientific and medical communities relying on accurate citations). The widespread infiltration of fake references into peer-reviewed literature and the failure of safeguards to detect them indicate a direct AI Incident rather than a mere hazard or complementary information. The harm is realized and ongoing, not just potential.

The event explicitly involves AI systems (LLMs) generating fake citations ('hallucinations') that have been incorporated into scientific papers, causing harm to the scientific community and potentially to public health through compromised clinical guidelines. The harm is realized and documented, not merely potential. The AI system's use in generating citations is the direct cause of the fabricated references. This constitutes a violation of intellectual property rights and scientific integrity, fitting the definition of an AI Incident. The event is not merely a warning or potential risk (AI Hazard), nor is it a response or update (Complementary Information).

The event explicitly involves AI systems (large language models) generating fabricated citations (hallucinations) that have been incorporated into published scientific literature. This has directly led to harm by contaminating the scientific record with false information, which can mislead researchers, clinicians, and policymakers, thereby violating intellectual property rights and harming communities dependent on accurate scientific knowledge. The studies document realized harm, not just potential risk, and highlight failures in existing safeguards. Hence, the event meets the criteria for an AI Incident rather than a hazard or complementary information.

The article explicitly identifies AI systems (generative AI tools) as the source of fabricated references and inaccurate content that have already appeared in published academic papers, books, and legal decisions. These AI hallucinations have directly led to harm by corrupting the scientific evidence chain, misleading legal outcomes, and disseminating false information, which affects health decisions, legal fairness, and societal trust. The harms are realized and documented with data and examples, meeting the criteria for an AI Incident. The article also discusses the systemic risk and challenges in remediation, reinforcing the significance of the harm caused.

The article explicitly discusses AI systems generating fabricated references (hallucinations) that have been incorporated into academic papers, leading to misinformation and undermining trust in scientific research. This constitutes a violation of intellectual property rights and harms the academic community and society by spreading false information. The harm is realized, not just potential, as thousands of papers with fake references have been published and remain unretracted. Therefore, this qualifies as an AI Incident due to direct harm caused by AI system use and malfunction in academic research.

The article explicitly identifies AI systems (AI tools used for research assistance, text generation, and citation generation) as the source of fabricated references and inaccurate content. These AI hallucinations have directly led to harm by corrupting scientific literature, which in turn can affect clinical guidelines and patient care, as well as legal decisions and journalistic integrity. The harm is realized and ongoing, not merely potential. Therefore, this event qualifies as an AI Incident because the AI system's use has directly led to violations of the integrity of knowledge (harm to communities and potentially harm to health through compromised medical guidelines) and breaches trust in expert work. The article also highlights the systemic nature of the problem and the difficulty in remediation, reinforcing the significance of the harm.

The article explicitly mentions AI-assisted drafting inserting fabricated sources into biomedical papers, with a significant number of fake citations found across millions of papers. This misuse of AI has directly led to the dissemination of false information in medical literature, which can harm patient care and the integrity of scientific evidence. The harm is indirect but material, as false citations propagate through research and clinical practice. The event involves the use of AI systems in generating content that causes violations of trust and potential harm to health, fitting the definition of an AI Incident.

The article explicitly mentions the use of AI systems (LLMs) in generating professional reports that contain fabricated citations and quotes, which are false and misleading. This constitutes a violation of intellectual property rights and harms the credibility of professional services, thus meeting the criteria for harm to rights and communities. The harm is realized, not just potential, as evidenced by refunds paid and reports removed. The AI system's use and malfunction (hallucination) directly led to these harms. Hence, the event is classified as an AI Incident.