The capabilities of large-scale language model applications such as ChatGPT continue to attract attention. Healthcare is one area that can benefit from its capabilities.
Everything from writing term papers to writing legal briefs can benefit from AI chatbot applications. Unfortunately, these chatbots are ill-equipped to do many of the tasks they are given. Regular reports produced by ChatGPT can get failing marks due to poor organization, reasoning, and writing. Legal briefs have been rejected due to fabricated case law and citations.
However, the potential of AI chatbots is clear, and the pace of development, despite challenges, means that such tools will help many industries, especially those with conversations and those that filter out important information from conversations. means to Conversational AI in healthcare is no exception.
The Benefits of Conversational AI in Medicine
Healthcare is already highly digital. Clinics and hospitals are equipped with specialized equipment such as digital thermometers, MRI machines, and hematology analyzers. However, the instrument and its measurements do not tell the whole story. It can provide a large amount of data, but some data comes solely from conversations with patients and caregivers. This is where conversational AI will benefit healthcare and become preferable to traditional patient engagement. Here are some possibilities.
Medical history taking and examination. Conversational AI can, or will soon be, trained to take medical histories from patients, ask them about their symptoms and concerns, and record, transcribe, and summarize the results for doctors to read.
Create transcripts of conversations. AI chatbot technology can also complement other technologies, such as electronic medical records, in word-intensive medical situations, such as recording tests and procedures. In keeping these records, the technology helps to properly time patient visits and hand off patients from one doctor or nurse to another at the end of the day.
Automate repetitive tasks. While there are very different areas of AI that help doctors make diagnoses and identify treatment options, conversational AI holds promise in automation as well. These tools could handle many of the mechanical processes that doctors, nurses, and even pharmacists, for example, need to give instructions to patients. They will be able to repeat as well as paraphrase instructions as needed to patients without running out of patience like humans.
AI challenges in healthcare
The challenges of using conversational AI tools in healthcare are significant and need to be resolved before they are widely used.
AI chatbots that can be used for healthcare can be expensive to acquire and maintain.
Forgery. ChatGPT and other large language models can produce clearly false answers and output. Even more dangerous in a medical context, they can spew subtly false statements. If Tool claims the patient was not allergic to penicillin, it could be fatal if the opposite is true.
price. In many cases, conversational AI tools and the resources (such as data centers) required to operate them can be prohibitively expensive. AI chatbots that can be used for healthcare can be expensive to acquire and maintain.
privacy. Malicious attackers can hack conversational AI tools and expose private patient data and personally identifiable information. This data includes both the patient’s answers to the AI tool’s questions and the questions the patient asked her AI tool. For example, if a patient asked her AI her chatbot in the office to check aspects of her health record, that record would remain open to extraction hacks, leaving hospitals and pharmacies facing lawsuits and fines. You will be exposed to risk.
risk and uncertainty. Currently, there is no legal or regulatory framework to justify AI tools playing a significant autonomous role in healthcare. There is also a lack of standard insurance mechanisms to mitigate potential institutional risks for companies using such systems.
The healthcare industry expects conversational AI to play an increasing role in healthcare in the future, but we need regulation and governance policies to help address some of the challenges.
