Poll: 60% of adults are dissatisfied with AI in their healthcare

According to a survey conducted Pew Research Center.

A 2022 survey that included responses from over 11,000 American adults found that 66% of women said they would feel uncomfortable if their doctor relied on artificial intelligence in their care, while 54% of men expressed discomfort.

However, 38% of respondents said using AI to diagnose and treat patients would generally improve patient health, while 33% said it would lead to worse outcomes. Twenty-seven percent said it would make little difference.

Of the respondents who saw racial and ethnic bias in healthcare as a problem, 51% said it would improve if AI was used to diagnose diseases and recommend treatments, while 15% said its use would worsen the situation. A third of respondents said that the problem of bias and unfair treatment would remain about the same.

However, 57% of respondents said that the use of AI for disease diagnosis and treatment recommendations would worsen the patient-physician relationship. However, 40% said that artificial intelligence in healthcare would reduce healthcare provider errors.

Three-quarters of adults surveyed expressed concern that healthcare providers would adopt AI too quickly before fully understanding the risks to patients, and 37% said using AI would compromise the security of medical records.

In terms of specific use cases for AI, 79% of respondents said they would not want to use an AI-enabled chatbot if seeking mental health support, with only 20% saying they would.

In a group of respondents who were aware of mental health chatbots prior to the survey, 71% of them said they would not want to use AI chatbots for their own mental health, and 46% of American adults said that AI chatbots intellect should only be used by people who are also seen by the therapist. Twenty-eight percent said they should not be available to people at all.

WHY IS IT IMPORTANT

Investing in AI tools for healthcare increased after the COVID-19 pandemic.

Still, clinicians and patients need to trust the performance and accuracy of machine learning models and be sure that AI is not biased.