Researchers at the National University of Singapore (NUS) have launched new artificial intelligence (AI) programmes with the dual aim of providing Asia with a much-needed boost in terms of AI research while also helping tackle societal problems.
The Centre on AI Technology for Humankind, or AiTH (pronounced “eighth”) was established in January by the university’s business school and had its official − albeit virtual − launch in April.
That same month, the university’s N.1 Institute for Health presented iDentif.AI (pronounced “identify”), an AI platform that can help identify drug combinations for infectious diseases.
David de Cremer, AiTH’s director and a business professor, told Times Higher Education that “Western universities and companies have opened centres focusing on the ethics and responsibilities of AI, but Asia was lagging behind a bit”.
“Asia is the number one adopter of the technology, but Asian companies engage in less data sharing because of low trust,” he said. “It’s using the technology more boldly, more quickly. But there’s a big discrepancy in the way data is used in, say, China and Western countries.”
Professor De Cremer, who moved from the University of Cambridge to NUS about a year ago, said that Singapore was an ideal place to explore such issues. “The theme of the centre needs a global approach, and at the same time needs to be introduced more fiercely in Asia. Hence Singapore, as the place where East meets West, is most appropriate.”
A report on Asia’s AI agenda by MIT Technology Review Insights in conjunction with NUS and other sponsors, surveyed more than 1,000 senior executives and showed that nearly 96 per cent of Asian respondents used AI in 2019, compared with 85 per cent in other parts of the world.
“Technology can be used for good or for bad; it needs to be regulated,” said Professor De Cremer, who explores the relationship between technology and humans in his upcoming book, Leadership by Algorithm: Who Leads and Who Follows in the AI Era?
“Working at a university, you think about your field in light of how it contributes to society,” he said. “There’s a responsibility that relates to being an academic. You put different ideas together; you theorise; you research.”
“But for corporations, it’s hard for them to step outside themselves,” he added. “Being a professor in a business school puts me in the middle. I teach a lot of executive MBAs who express to me their concerns about technology.”
The Covid-19 pandemic has amplified both the need for AI in medical research and the need to use data responsibly. Professor De Cremer gave an example of TraceTogether, a Singaporean app being use to track the disease, asking what will happen to the data gathered from that app after the pandemic is over.
Dean Ho, director of NUS’s N.1 Institute for Health, was working on the iDentif.AI platform when he heard about the novel coronavirus from China. iDentif.AI uses artificial intelligence to narrow down the number of experiments needed to discover ideal drug combinations and dosages.
“Globally, it was an interesting time,” Professor Ho said. “Our work was mostly in oncology, but I did feel that infectious diseases required more urgent attention. If it’s a serious case, a patient can pass away very quickly.”
Professor Ho’s research team chose 12 drug candidates for treating a type of viral lung infection. “If you have 12 drugs and 10 different dosages, that makes a trillion possible combinations. Traditional approaches cannot scan that many combinations. But we can drastically shrink down the number of experiments needed.”
A paper introducing the platform was published in Advanced Therapeutics in April. Since that first round of research, his team have shifted to working specifically on Covid-19.
Professor Ho has taken a distinctly multidisciplinary approach to his field of biomedical engineering, collaborating with economists and others in the corporate sector on his research. His paper is co-authored with academics from Shanghai Jiao Tong University and Fudan University, but also with experts from EpiPoint, MRIGlobal and consulting firm KPMG.
“I wanted independent contributors to help us assess what it would take to get this technology deployed and to sustain it. It’s rare, in a technical paper, to have several pages of non-technical discussions,” he said. “But technology alone cannot change health care.”
Professor Ho does not want to commercialise iDentif.AI. “We don’t want to charge patients to use IDentif.AI," he said. "We want this to be for the public benefit. It we can prove that this platform can help keep patients out of the ICU, or get patients out of the ICU faster, it may be a cost-neutral solution.”
AI can help ensure that medical systems are prepared for the next potential pandemic, he added: “If we encounter a Covid-like situation again, instead of using traditional drug combinations for months, we may be able to rapidly identify more potent combinations from the start.”
