SINGAPORE: Despite the Ministry of Health’s S$200 million (US$155 million) plan to bring AI into public healthcare over five years, many independent GPs still face a simple choice — pay out of pocket or skip it altogether.
“Honestly, it’s about cost, access, and whether it’s really worth it,” Dr Song Majinyang says. She uses ChatGPT’s free trial to draft emails, workflows, and standard operating procedures, easing the day-to-day grind. But when it comes to paying for AI tools that could help with patient care, she’s holding back.
AI comes with a price tag
According to CNA, two out of five GPs embraced AI for clinical work. The others stick to using it for basic admin tasks—or avoid it entirely. For those who adopt AI, the costs often come straight from their own pockets.
For tinier health centers, spending for AI isn’t just a question of money. Dr Song worries that commercial tools like ChatGPT give generic answers, while more advanced clinical AI is either tied up in pilot programs or simply hard to access. On top of that, patient confidentiality remains a critical concern. “Doctors need to be sure that patient data stays secure, and no one else is reading behind the content,” she says.
RadLink Group, a diagnostic imaging provider, has tackled this by keeping all imaging data local. Dr Eng Chee Way, the company’s medical director, explains that AI helps with chest X-rays, mammograms, and CT lung scans—but nothing leaves their servers, ensuring patient information never goes to the cloud.
Different views on AI’s value
Dr Roland Xu of Procare GP Clinic, who hasn’t adopted any AI tools, believes the technology is better suited for hospitals managing complex cases. “If patients don’t even ask whether AI is being used, I don’t see the immediate need for us to adopt it,” he says.
Others are more optimistic. Dr Koay points out that AI can catch early warning signs in cancer screening, even when scans look normal, helping doctors to intervene sooner. Dr Eng, a strong supporter of AI, emphasizes collaboration: AI is tireless and consistent, but it can miss details, making the combination of human and machine the “best outcome”—like a pilot and autopilot working together on a plane.
As technology improves and its potential to prevent illness grows, more clinics may start to see it as an essential part of modern medical practice.
