Recent advances in large language models, such as OpenAI's ChatGPT, have implications for scholars and medical journals. These AI algorithms can generate new content, but their outputs are highly error-prone and prone to fabricating information. The International Committee of Medical Journal Editors (ICMJE) has updated its recommendations for scholarly work published in medical journals to address the use of AI-assisted technologies. The ICMJE now requires authors to disclose any use of AI-assisted technologies in their manuscripts and assume full responsibility for the accuracy and integrity of the content. AI-assisted technologies should not be named as authors or cited as primary sources. Additionally, peer reviewers are prohibited from using AI tools that may compromise manuscript confidentiality. The development and use of AI-assisted technologies must proceed with transparency and caution, as they can generate misinformation and error.
Mon, 24 Jul 2023 04:31:04 GMT | CMAJ