AI, ChatGPT Poised to Potentially Transform Healthcare, Science

According to a new report from Accenture, the growth of artificial intelligence (AI) and ChatGPT is leading to transformations in science, business, healthcare, and society, positively impacting human creativity and productivity.

With advancements in modern technology, researchers are continuing to develop various types of AI, using it to transform different systems and processes. According to the report, 98 percent of global executives agreed that in the next three to five years, their organizations' strategies would involve AI foundation models.

Amid the growing prominence of AI, ChatGPT is a service that continues to gain recognition for performing various types of tasks within several industries, one of which is healthcare. ChatGPT is a natural language processing tool driven by AI.

In February, researchers from Massachusetts General Hospital (MGH) and AnsibleHealth found that ChatGPT can pass the United States Medical Licensing Exam (USMLE). Along with this, their research indicated that the tool could impact medical education.

According to the report from Accenture, 42 percent of companies expressed interest in making a large ChatGPT investment in 2023. Solutions such as this have the potential to transform industry practices. However, there are certain steps that can maximize the abilities of AI. These include maintaining a business-driven goal, prioritizing people, preparing enough proprietary data, and making appropriate investments in technology infrastructure, architecture, operating model, and governance structure.

However, the growth of ChatGPT also brings forth questions surrounding responsibility in using AI, specifically regarding legal, ethical, or reputation risks that may be involved. The report suggests building risk assessment controls at the design stage and implementing responsible AI principles throughout the business.

The report concluded that reinventing work to allow AI to gain a valuable place would be in a company's best interest.

The use of AI and ChatGPT in healthcare continues to grow as researchers discover the various ways it can provide support.

For example, researchers wrote in Cureus in March that ChatGPT could answer first-order and second-order questions in line with the competency-based medical education (CBME) curriculum for microbiology.

Researchers noted that a major reason for this study was due to limited evidence related to the abilities of ChatGPT to answer first- and second-order knowledge questions. This study involved authors preparing six first- and six second-order knowledge questions according to the eight modules. They defined first-order knowledge questions as more straightforward and second-order knowledge questions as more complex.

Following reviews for content validity and insertion into ChatGPT, they found that ChatGPT reached an overall accuracy of 80 percent with limited differences in its ability to answer the two types of questions. These findings led them to conclude that ChatGPT could answer both types of questions related to microbiology.

Another study from March found that ChatGPT provided accurate information when questioned about potential cancer myths and misconceptions.

Researchers pulled questions from the National Cancer Institute's (NCI) “Common Cancer Myths and Misconceptions” web page to guide this study, comparing NCI answers to those delivered by ChatGPT.

Following review, they found that the percentage of overall agreement for accuracy was 100 percent for NCI answers and 96.9 percent for ChatGPT answers. Despite this high percentage, researchers noted that ChatGPT language was often unclear, meaning that users could incorrectly interpret information.