Innovations in the field of artificial intelligence (AI) have created an avenue for expediting the production of specialized health software. As studies reveal, generative AI (GenAI) can significantly accelerate the development of healthcare tools designed for disease prevention, such as those for diabetes.
One notable study was published online on March 6 in the Journal of Medical Internet Research. Generative AI, predictive technology that anticipates the successive word in a sentence, streamlines the design process for these types of software. In this context, the AI takes on the form of a 'chatbot' that is capable of generating coherent and logical responses to inquiries. Essentially, generative AI tools such as ChatGPT can succinctly summarize complex texts.
The research spearheaded by NYU Langone Health delves into the practical application of ChatGPT in the development of a software program that encourages diabetes patients to maintain healthier lifestyles by sending text messages. The team wanted to confirm if the integration of AI could speed up the creation of a personalized and automatic messaging system (PAMS).
In the context of their research, eleven evaluators from various fields such as medicine and computer science were able to create a version of a diabetes tool in 40 hours. This is a stark contrast to the previous attempt where no AI assistance was used, taking more than 200 programmer hours.
The researchers concluded that ChatGPT helps facilitate more effective communication between technical and non-technical team members. In turn, this fast-tracks the design and creation of computational solutions to medical problems. If this methodology can be efficiently scaled, this could lead to radical changes in the healthcare software design landscape.
The GenAI tools used in the study exemplify sensitivity. Subtle shifts in how a question is asked may produce divergent answers. This calls for a skill known as 'prompt engineering', which encapsulates both intuition and experimentation. Doctors and nurses, given their understanding of nuanced medical contexts, can deliver strategic prompts that aid in fostering better communication with engineers, making them valuable team members in this process.
This improved line of communication was demonstrated in the study. Clinical members were able to input their ideas in plain English into the ChatGPT tool, which then translated the entries into technical language that the team's software engineers could easily understand and translate into code. While AI technology was instrumental in the design process, human developers were still integral for the final stages of code generation.
The research suggests that tools like ChatGPT can help democratize the design process of healthcare software, enabling doctors and nurses to have a direct hand in its creation. GenAI-assisted development promises to deliver computational tools that not only meet the highest coding standards, but are also reliable and user-friendly.
Disclaimer: The above article was written with the assistance of AI. The original sources can be found on ScienceDaily.
