Recently, OpenAI collaborated with WHOOP to launch a personalized health and fitness coach powered by GPT-4. WHOOP Coach solves a variety of fitness and health-related questions.
It can, for example, answer questions like “What was my lowest resting heart rate ever?” a “What weekly workout schedule would help me reach my goal?” — all while providing tailored advice based on each individual’s unique body and aspirations.
In addition to WHOOP, OpenAI has collaborated with Summer Health, a text-based pediatric care service that uses GPT-4 to aid its clinicians. Summer Health has developed and introduced a new tool that uses GPT-4 to generate visit notes automatically from a doctor’s thorough written remarks.
The pediatrician rapidly reviews these notes before sharing them with parents. Summer Health carefully fine-tuned the model in conjunction with OpenAI, implemented a clinical review procedure to verify accuracy and relevance in medical situations, and continues to enhance the model based on expert feedback.
Furthermore, GPT Vision has found uses in radiography. Microsoft has published a paper titled ‘Exploring the Boundaries of GPT-4 in Radiology’ that evaluates the performance of GPT-4 in text-based applications for radiology reports.
GPT-4’s capacity to process and comprehend medical pictures ranging from X-rays to MRIs is one of its key applications in radiology. According to the research, “The radiology report summaries created by GPT-4 are comparable, and in some cases, even preferred over those written by experienced radiologists.”
Furthermore, Be My Eyes is leveraging GPT-4’s multimodal features, notably the visual input aspect, to improve its virtual assistant software. Be My Eyes assists visually challenged people with skills such as item identification, text reading, and navigation.”
On the mental wellness front, many people have tried ChatGPT as a therapist. Many people have found ChatGPT to be beneficial in terms of practical guidance and human-like interaction, providing a unique alternative for those who are unable or reluctant to undergo professional treatment.
What Others Are Doing
Before OpenAI, both Google and Apple made substantial advances in the healthcare field employing LLMs.
Google has announced MedLM, a set of foundation models tailored to certain healthcare use cases. There are now two MedLM models, both built on Med-PaLM 2, that provide flexibility and suit the different demands of healthcare organizations.
Apple, on the other hand, intends to include more health detection functions in their forthcoming line of watches, focusing on illnesses such as hypertension and apnea, among others.
Furthermore, Isomorphic Labs, a London-based spin-off of Google AI R&D division DeepMind, has formed strategic alliances with two pharmaceutical behemoths, Eli Lilly and Novartis, to use AI to develop novel pharmaceuticals to treat ailments.
Oracle released Clinical Digital Assistant last year, which interfaces with Oracle’s EHR systems and allows clinicians to use voice commands for operations such as note-taking, appointment scheduling, and drug ordering.
Collaboration with healthcare institutions and researchers, similar to Isomorphic Labs, would provide OpenAI with access to diverse medical data, which is critical for improving AI models and expediting product development.
Furthermore, diversifying OpenAI’s revenue streams reduces exposure to swings in other industries.
