Technological innovations are significantly impacting our daily lives and changing the way we obtain information. This is also the case for health-related information. In this context, innovative artificial intelligence (AI) systems are emerging to offer changes in healthcare, including diagnostic support, treatment optimization and personalized health management. However, studies to date suggest that not all individuals are willing to use such AI-enabled health services.
In response, IRTIS researchers conducted research that originated from the SYRI (National Institute for Research on the Socioeconomic Impact of Disease and Systemic Risk) project. Data collection took place in October 2023 and involved a total of 4,775 internet users aged 18 to 95. The results showed that almost a third of the sample of Czech adults (28%) had used an AI system such as ChatGPT or Bard in the past year. Younger respondents, men and people with a university or higher professional education used AI slightly more than other groups.
Respondents used AI for a wide range of activities. "It turns out that Czech adults used AI to find tips for healthy recipes, for proofreading texts, or to assist them with programming. Parents, for instance, used ChatGPT to generate tales for their children. Testing the AI's skills and knowledge was no exception, as people tried to catch the system making mistakes," explains Natálie Terčová, a researcher from the IRTIS Institute, a research department at Masaryk University (irtis.muni.cz).
About a third to a half of people who have used AI in the past year would be willing to use it to search for health information, depending on the type of information. Most people would use AI to search for information about healthy eating (48%), and the least would use it to diagnose an illness (32%). "One possible reason for the reluctance to use AI to search for information about illnesses may be concerns about protecting one's privacy and sensitive health data," Terčová says. The research also showed that although people are willing to use AI to search for health information, it doesn't automatically mean they trust that information. For example, only 27% of those who " rather agree" that they would use AI to look up information about a disease diagnosis would trust these recommendations. Conversely, more than half of those who have not used AI before said they would not trust AI-generated recommendations. Distrust of AI health recommendations also increased slightly with age.
"In this research, we focused on the positive aspects of using AI in the context of health, and deliberately did not address the potential risks. However, the relatively low trust in health information provided by AI also highlights the possible risks, which include the current low credibility and frequent inaccuracy of the information provided," adds Prof. David Šmahel from IRTIS. Only the future development of AI will show how useful AI technologies will be in the field of health and health information retrieval.
