SINGAPORE - Artificial intelligence (AI) technologies can bring a wide range of opportunities and benefits, but they also come with the risk of perpetuating biases and inequitable distribution of those benefits.
Much needs to be done to ensure AI development avoids such pitfalls and progresses in a way that delivers social good, said experts on a panel at the Asia Tech X Singapore event on Wednesday (July 14).
The role of ethics in AI systems has been "late to the game" and is often brought up only after problems surface and AI systems begin "automating inequality", said Singapore University of Social Sciences (SUSS) ethics professor Jennifer Ang.
Instead of putting the technology first and asking what AI can do, ethics should be introduced at the beginning of the development process and not at the end as an afterthought, said Associate Professor Ang, whose research includes a focus on ethics in AI.
"We need to first define what social good is, rather than what AI is capable of. Once we start with the idea of what social good is and we all agree to it, AI should develop to meet those goals rather than the other way around."
Mr Royce Wee, head of global public policy at Alibaba Group in Singapore, highlighted the role of government regulations in defining the boundaries of AI use. He cited the recent introduction of several key pieces of legislation in China, including the new Data Security Law, which comes into force in September and will regulate a wide range of issues related to data use in the private sector.
Alibaba is investing heavily in AI technology like chatbots, machine translation, user insights, marketing algorithms and logistics management. While this is partly a business necessity due to strong competition from rivals both within China and overseas, Mr Wee said the e-commerce giant's development of AI also aims to contribute to solutions for society's problems.
He cited how Alibaba's research and development facility created diagnostic tools during the Covid-19 pandemic to quickly and accurately determine whether a patient was infected with the virus and how severe the disease was likely to become, based on CT scans.
The other speakers on the virtual panel were Mr Omar Sultan Al Olama, Minister of State for Artificial Intelligence in the United Arab Emirates; Mr Raymund Enriquez Liboro, the commissioner and chairman of the National Privacy Commission in the Philippines; and Dr Wendell Wallach, the Emeritus Chair of Technology and Ethics Studies at the Yale Interdisciplinary Centre for Bioethics.
Mr Omar spoke about the need for governments to balance the interests of the public with those of private, profit-seeking companies, which may deploy AI technology without a strong understanding of the ethical implications.
Mr Liboro said organisations that implement AI technologies should always try to anticipate and mitigate potential risks, and work closely with other stakeholders to resolve problems that do arise. They should also clearly define who holds responsibility for the operations and effects of AI systems and how any concerns can be raised, he said.
Dr Wallach warned that AI acts as an amplifier of other technological trends and can worsen disruptions and income inequality brought about by the digital economy. He said measures, such as the Group of Seven (G-7) nations' proposal for a global minimum corporate tax rate of at least 15 per cent, are a good first step, but will require cooperation between governments.
The panel was moderated by the National University of Singapore's (NUS) law dean, Professor Simon Chesterman, who is also senior director of AI governance at AI Singapore, a national programme.
The Asia Tech x Singapore (ATxSG) AI Conference, organised by the Infocomm Media Development Authority, brought together industry leaders, policymakers and experts to discuss leading technological trends.
The four-day event ends on Friday.