The advent of artificial intelligence (AI) can help solve the growing healthcare challenge of aging populations, which many nations are battling, though the usual challenges remain.
According to Singapore’s Health Minister Ong Ye Kung, three significant trends are converging to make this an unprecedented time for tackling healthcare challenges: AI, genomics, and a shift in focus toward preventive healthcare.
Also: How digital twins could save time, money, and lives in developing prescription drugs
Genomics, or precision medicine, tailors healthcare to a person’s unique genes and environment, allowing for more accurate treatment or preventive measures. At the same time, generative AI (gen AI) opens up opportunities to enhance the diagnosis and treatment of diseases, Ong said at a fireside chat during ST Engineering’s InnoTech.Healthcare conference this week.
Per the World Health Organization, one in six people worldwide will be aged 60 years and above by 2030 — accounting for 1.4 billion of the global population. By 2050, that number will double to 2.1 billion, and the number of people aged 80 years and older is projected to reach 426 million.
By 2050, two-thirds of the population aged 60 and above will also live in low- and middle-income countries.
Ong noted that aging drives several developments in healthcare because it correlates with higher disease load and illnesses and contributes to escalating healthcare costs, not just for patients but also governments and nations.
Also: Hooking up generative AI to medical data improved usefulness for doctors
Countries like Singapore are working to ensure they can cope with their aging populations. Since aging is inevitable, Ong thinks the country should focus on preventive measures and tapping technology.
Data coupled with AI, for instance, can provide valuable insights into which diseases certain individuals are susceptible to and provide the necessary preventive care, Ong said. He added that data should be stripped of personally identifiable information (PII) and anonymized.
<!–>
Singapore is currently working to ensure it has the infrastructure in place to support such initiatives, including a common electronic medical record (EMR) platform that all healthcare service providers, including hospitals and private clinics, can access. Ong said this will require GPs (general practitioners) to contribute patient data to the centralized data platform.
He noted that legislation must be in place to mandate such requirements while safeguarding the data.
Also: AI gold rush makes basic data security hygiene critical
Singapore has proposed a Health Information Bill that the country says provides the framework to “govern the safe collection, access, use, and sharing of health information” across its healthcare ecosystem. While the country’s EMR was established in 2011 as a centralized health data repository, it is mainly used by public healthcare institutions. Participation by private providers is voluntary – just 15% of them participated in its use as of October 2023, according to Singapore’s health ministry.
Ong said there is no single holistic profile of an individual’s health information, as data is fragmented and scattered across different providers. The health ministry says the national EHR system would only require key health information, such as diagnosis, medications, allergies, and laboratory reports. Providers would be allowed access to patients’ summary medical records that are relevant and required for them to provide care.
The proposed bill, which is scheduled to be read in parliament next week, would mandate that all licensed healthcare providers contribute data to the electronic health repository. It also lays out a set of cybersecurity and data security requirements that healthcare providers contributing to or accessing the data platform must meet. This includes reporting cybersecurity incidents and data breaches, such as unauthorized access.
Also: OpenAI’s GPT-4 can diagnose eye problems and suggest treatments like your doctor
Ong noted that robust data management is a key issue, and further safeguards will be necessary with the emergence of genomics and AI. For instance, he said that considering genomics information in hiring decisions will create ethical issues, adding that its use in underwriting life insurance policies has also been outlawed in some jurisdictions.
For Ong, public trust in the use of AI for healthcare will depend on there being controls in place to manage these concerns.
Addressing the deepfake problem
Deepfakes also popped up in discussions at the conference as an AI risk. Singapore’s Senior Minister of State in the Prime Minister’s Office Desmond Tan noted that deepfakes cannot be ignored as cybercriminals increasingly turn to advanced AI tools in their attacks.
During his speech at the conference’s main track, InnoTech.AI, Tan mentioned how scammers can create realistic deepfake videos with just a few photos and a short audio clip pulled from the internet or social media platforms.
Also: 80% of people think deepfakes will impact elections. Here are three ways you can prepare
He said deepfake content in Singapore climbed 500% last year compared to 2022, noting that such risks must be addressed so AI can be harnessed for its benefits, such as facilitating the early detection of diseases.
ST Engineering is hoping to help organizations do that with the launch of its Einstein.AI platform. Einstein.AI is designed to validate the authenticity of content by fact-checking and detecting audio and video deepfakes.
The software analyzes the transcript of the content and fact-checks the information against established news platforms or any other media agencies that customers choose. It also runs stance and sentiment analyses to identify any potential hate speech in the content.
ST Engineering said that organizations running the Einstein.AI platform can potentially set parameters for these checks.
Also: No needles necessary: This wearable can track your cholesterol and lactate levels
The AI platform does not conduct facial recognition and, therefore, cannot authenticate the identity of an individual. Rather, it searches for patterns to help determine if the content itself is synthetic or has been manipulated.
Developed in-house, ST Engineering trained Einstein.AI by feeding it video and audio clips of individuals with corresponding deep fake twins to identify the nuances between them.
ST Engineering is currently submitting research proposals to partner with higher education institutes on research initiatives to improve its deepfake detection capabilities, which can then be embedded into Einstein.AI.
The vendor added that it is also open to tapping other AI models via industry partnerships to further supplement the platform.
