This study explores patient perspectives on informed consent for medical AI, focusing on whether doctors should disclose the use of AI tools in diagnosis and what information should be provided. A survey experiment with 1000 South Korean respondents was conducted to assess the perceived importance of AI-related information in treatment decisions. The study found that the use of AI increases the perceived importance of information related to its use compared to when a human radiologist is consulted. Participants perceived AI-related information as more important than or similar to regularly disclosed short-term effects information. Gender, age, and income significantly influenced the perceived importance of AI information. The study supports the disclosure of AI use in diagnosis during informed consent. However, disclosure should be tailored to individual patient needs, as preferences vary by gender, age, and income. Ethical guidelines for informed consent when using AI in diagnoses beyond legal requirements are recommended. The study highlights that patients value information about AI use more when it is involved in diagnosis compared to when a radiologist is used. Patients also perceived AI information as more important when the recommended treatment posed higher risk. However, the prevalence of AI use and its performance relative to human radiologists did not significantly affect perceived importance. Socio-demographic factors such as gender, age, and income significantly influenced the perceived importance of AI information. The study concludes that patient preferences for AI-related information differ by socio-demographic factors. A personalized informed consent process is necessary to accommodate individual information needs. The findings suggest that AI-related information should be tailored to patients' preferences, and that future research should explore how newly introduced AI risks and the role of AI in decision-making affect patient information needs. The study also emphasizes the need for clear guidelines and training for healthcare professionals to ensure effective AI disclosures.This study explores patient perspectives on informed consent for medical AI, focusing on whether doctors should disclose the use of AI tools in diagnosis and what information should be provided. A survey experiment with 1000 South Korean respondents was conducted to assess the perceived importance of AI-related information in treatment decisions. The study found that the use of AI increases the perceived importance of information related to its use compared to when a human radiologist is consulted. Participants perceived AI-related information as more important than or similar to regularly disclosed short-term effects information. Gender, age, and income significantly influenced the perceived importance of AI information. The study supports the disclosure of AI use in diagnosis during informed consent. However, disclosure should be tailored to individual patient needs, as preferences vary by gender, age, and income. Ethical guidelines for informed consent when using AI in diagnoses beyond legal requirements are recommended. The study highlights that patients value information about AI use more when it is involved in diagnosis compared to when a radiologist is used. Patients also perceived AI information as more important when the recommended treatment posed higher risk. However, the prevalence of AI use and its performance relative to human radiologists did not significantly affect perceived importance. Socio-demographic factors such as gender, age, and income significantly influenced the perceived importance of AI information. The study concludes that patient preferences for AI-related information differ by socio-demographic factors. A personalized informed consent process is necessary to accommodate individual information needs. The findings suggest that AI-related information should be tailored to patients' preferences, and that future research should explore how newly introduced AI risks and the role of AI in decision-making affect patient information needs. The study also emphasizes the need for clear guidelines and training for healthcare professionals to ensure effective AI disclosures.