Physicians are excited but cautious about the use of artificial and augmented intelligence in medicine, according to a survey published by the American Medical Association on Thursday.
AMA surveyed more than 1,000 doctors for their thoughts on AI, which the advocacy group defines as augmented intelligence, a type of AI technology that still requires human involvement. The survey revealed the majority of respondents see advantages to AI in healthcare but some are concerned over its potential effect on patient relationships and data privacy.
Read more: HHS finalizes AI transparency rule
In June, AMA delegates voted to study the benefits and unforeseen circumstances of AI, including large language models such as GPTs and other intelligence-generated medical content, and propose appropriate state and federal regulations. The survey, which was conducted in August, was meant to understand physician sentiment on AI and identify potential use cases for the technology.
Here are five takeaways from the survey:
1. Docs see the potential AI
Nearly two in three respondents said AI is positively affecting patient care. Only 11% cited disadvantages in using AI for care. Those trends were pretty universal among different types of physicians. AMA found no significant difference in sentiment between primary care providers and specialists. Only a minor difference was found between employed physicians and those who own their own practice, with the latter slightly more skeptical of AI.
2. A big plus seen for diagnostic assistance and automating prior authorizations
The survey found physicians envision AI tools assisting them with clinical diagnosis as well reducing their administrative burdens by improving documentation and automating prior authorizations. Improving documentation was pegged by 74% as an area where AI could be relevant to their work, with 69% saying the same for automating prior authorizations. Around 72% of respondents say AI can help improve their diagnostic ability. Companies such as Microsoft subsidiary Nuance and big tech firm Amazon are developing AI-enabled clinical documentation software.
3. A gap in sentiment
Respondents remain at odds over whether to be excited or cautious about AI in healthcare. Around 41% of survey respondents were equally concerned and excited about its use in their professional lives, while 30% reported feeling more excited than concerned -- nearly the same number felt the opposite. The survey found respondents were most worried about AI harming patient privacy and patient-physician relationships. Eighty-six percent of physicians reported a desire to be either responsible for or consulted before their practices adopt tools that use AI.
4. Most are still not using AI in practice
While 62% of respondents said they have not incorporated AI-use cases into their practices, AMA president Dr. Jesse Ehrenfeld said momentum for it is evident in certain areas. “It is reaching a growing number of practices,” Ehrenfeld said. “But it's certainly not integrated daily into the majority of clinicians practices.” The survey found transcription and translation services and draft response generation technologies were the areas where AI adoption was most likely to happen within the year.
5. Or ChatGPT
Most physicians aren’t flocking to OpenAI's publicly available generative AI model ChatGPT, which celebrated its one year anniversary in November. Around 62% of respondents reported never used ChatGPT in any capacity and 72% said they have never used it for work.
