Generative AI refers to AI that generates results based on the user's specific needs. It generates various contents such as novels, poems, images, videos, coding, and art by learning from data sources.

A typical example is ChatGPT, which provides detailed answers just by simply jotting down questions.

In the medical field, generative AI is receiving attention to improve doctors' efficiency and increase the accuracy of medical treatment.

Most recently, Samsung Electronics subsidiary Harman has entered the fast-growing medical generative AI market.

Harman said it has launched HealthGPT, a generative AI solution that enables doctors, researchers, and organizations to provide advanced patient care, conduct medical research, and make informed decisions last Friday.

The company said that HealthGPT is built on responsible AI principles.

"It has been extensively tested using an automated LLM (personal language model) testing framework and validated by healthcare experts," the company said. "HealthGPT provides real-time, context-aware clinical insights to help healthcare professionals make critical decisions.

It also utilizes an LLM fine-tuning framework to deliver optimal performance, it added.

Other features include an automated LLM evaluation framework, cost optimization using advanced deployment techniques such as quantization, and acceleration of drug discovery and research through data insights.

"With our AI expertise and ability to develop effective LLMs, Harman will help our clients fully capitalize on generative AI," said Nick Parrotta, President of Digital Transformation Solutions & Chief Digital and Information Officer at Harman.

Harman is one of many companies that have jumped into developing a generative AI with a focus on the medical field.

Earlier this year, Google also launched Med-PaLM, a chatbot designed to generate helpful answers to doctors and patient questions.

Med-PaLM has been rolling out to select healthcare organizations, including the Mayo Clinic in the U.S., since April of this year.

Google is also currently adding a "Multi-Modal" feature to help improve patient outcomes by synthesizing information such as x-rays and mammograms.

Amazon also launched AWS Health Scribe in August, a solution that uses generative AI and voice recognition to fill out patient medical records automatically.

The service, launched by its cloud division Amazon Web Services (AWS), recognizes conversations between doctors and patients and creates medical records in real-time.

Specifically, the service complies with U.S. healthcare laws and creates patient records but does not keep customer information. Amazon expects the tool to save time by generating specific information, such as medical terms and medications.

Microsoft also launched a medical record application called Docs Express in March, using Nuance, an artificial intelligence company it acquired in 2021.

The application transcribes conversations between doctors and patients in real-time and generates medical records using GPT-4, a multimodal LMM developed by OpenAI.

Nuance emphasized that Doctors Express saves seven minutes of time per patient and allows doctors to focus more on the patient.

Korean companies also dive into developing generative AI for medical purposes

While global companies are advancing technology one after another, there are also active movements in Korea.

For example, Kakao Brain has developed an AI CAD that generates draft readings for radiologists using large-scale models and generative models.

AI CAD is a solution that allows AI to write a doctor's opinion on the symptoms of a patient's chest X-ray image.

After recognizing that radiologists spend more time writing image readings than analyzing images, the company decided to develop such a system, and expects that their solution will help improve their work efficiency.

Naver has also previously introduced various AI technologies specialized in the medical field.

These include "Smart Survey," which automatically converts medical information into medical terms, "VOICE EMR," which creates medical records through voice recognition, "Patient Summary," which queries and analyzes medical records and checkup results, and "Clova CareCall," an AI call service that check up on users.

Challenges remain in using generative AI in Korea

However, there is still a challenge to standardize and quality control data from numerous medical institutions to create a good AI model.

This was recently highlighted at the K-Healthcare Business Forum, co-hosted by the Korea Biomedical Review (KBR), along with its sister publication, The Korean Doctors' Weekly, and the Korea Healthcare Industry Promotion Agency (KHIDI).

During the conference, Kakao Healthcare CEO Hwang Hee mentioned the importance of securing health data.

He noted that if obtaining data from medical institutions proves challenging, it becomes crucial to think about building an internal data analysis system within these institutions.

"This approach enables the processing of medical-related data internally rather than relying on external sources," he said.

There is also the fact that experts are pessimistic that generative AI will have a significant impact on the medical field in the near future, given Korea's unique reimbursement system.

"It is much cheaper to have a resident do the work of a generative AI in Korea," a doctor working at a tertiary general hospital told Korea Biomedical Review, asking to remain anonymous due to the sensitivity of the issue. "As a result, I am doubtful that hospitals will spend a lot of money to adopt the system."

For example, if it costs 1,000 won to develop a generative AI, the developer needs to earn about 3,000 to 4,000 won to gain profit and continue to update the software, but in Korea, the price will be much lower than that, he added.

The doctor stressed that this is a demoralizing factor for developers who have worked hard to develop a generative AI.

"For doctors working in the clinical field, many are questioning the rationale behind whether they should study and use something new at a time when they already have a heavy workload."

Considering these points, the doctor stressed that he does not believe that generative AI will have a big impact on the medical field in Korea in the foreseeable future.