Artificial intelligence (AI) is working its way into all sorts of technologies. Even healthcare is starting to feel it. For example, AI is being explored as a way to help predict the likelihood of future illness based on a patient's history compared to reams of data from other patients and statistical models. As a psychiatrist, how do you feel about AI?
There is a move to introduce the principles of the AI to psychiatry via technologies that would make it easier for psychiatrists to do their jobs. The pace is slow given the complexity of psychiatry as a medical discipline. But there probably is no stopping AI from eventually playing a role in daily practice. The only question is how big that role will be.
A recent survey conducted via a partnership between Sermo and researchers at Harvard and Duke Universities reveals just what psychiatrists think about AI. Some of the numbers might surprise you. Needless to say that the majority of psychiatrists surveyed do not think AI will ever completely replace them.
Key Survey Numbers
The Psychiatric Times reports that the survey solicited input from nearly 800 psychiatrists in 22 countries. Respondents were asked to assess the ability of AI to eventually replace human psychiatrists for completing complex tasks. Respondents were also asked about the risks and benefits of technology in their daily work.
Here are some of the more striking numbers:
- 17% said that technology will probably replace humans in providing empathetic care
- 67% said that technology would not make psychiatrists obsolete for complex tasks
- 54% said that future technology would be used to synthesize patient information for diagnostic purposes
- 75% said future technology would be effective for managing documents and medical records.
It is clear from the survey results that psychiatrists see AI and other technologies as beneficial to handling some of the more mundane tasks of daily practice. They do not think the technology will ever replace them completely.
AI Has Its Limits
It is easy to see how AI could be beneficial for certain kinds of mundane tasks. For example, an AI system capable of synthesizing patient data would make the lives of locum tenens psychiatrists a lot easier. Locum clinicians wouldn't have to comb through months and months of records to figure out where a particular patient is at. They could rely on AI software to present an accurate picture.
As exciting as that might be, AI does have its limits. The one thing AI cannot do – and probably will never be able to do – is think for itself. Indeed, the phrase 'artificial intelligence' really isn't accurate in describing what the technology does.
A truly intelligent machine would be able to identify what it doesn't know and then figure out ways to find the necessary information and learn from it. To date, there is no AI technology capable of doing that. AI technology can analyze all sorts of new information and then use it to reach conclusions, but it cannot identify what it lacks. It cannot figure out new ways to learn.
As such, there are plenty of things AI could never do in the field of psychiatry. Treating psychiatric patients is part science, part intuition, and part art form. AI is limited only to the scientific portion. Even at that, it has its limits.
There is little doubt that AI will transform psychiatry jobs in the future. As for its ability to eliminate the need for humans entirely, don't bet on it. No machine can do what the human brain can do.