Navigating AI’s role in Governance: Balancing Innovation, Ethics, and Public Trust

From left to right: Ms Eunice Huang, Head of AI and Emerging Tech Policy, Asia-Pacific, Google. Mr Jaime Ho, Editor, The Straits Times, SPH Media. Dr Janil Puthucheary, Senior Minister of State, Ministry of Digital Development and Information (MDDI), Singapore. Ms Sandy Kunvatanagarn, Head of Policy, APAC, OpenAI. Dr Carol Soon, Principal Research Fellow and Head, Society and Culture, Institute of Policy Studies, Lee Kuan Yew School of Public Policy.

Rapid advances in artificial intelligence (AI) are forcing Singapore’s policymakers to build regulatory frameworks and policy responses around a readiness to pivot as the technology evolves.

Singapore’s policy response to the evolution of AI technologies was among the key points discussed in a recent session at the Lee Kuan Yew School of Public Policy's (LKYSPP’s) Festival of Ideas 2024 involving leading thinkers and policymakers, including Senior Minister of State, Ministry of Digital Development and Information (MDDI), Dr Janil Puthucheary.

Panellists discussed the implications of changes in artificial intelligence technology on governance and society and focused on the immediate imperative of harnessing AI's potential while upholding ethical standards and public trust.

Chairperson of the session Dr Carol Soon, Principal Research Fellow and Head of Society and Culture at the Institute of Policy Studies, LKYSPP, opened the session with a comparison of how AI has developed from the 1960s using the example of Eliza, an early chatbot, to its phenomenal presence today across various domains including business, creative media and media industries, education, and work.

“With increasingly widespread adoption of AI technology by institutions and individuals, we are seeing benefits, but we are also witnessing some of the pitfalls … problems arising from the innovation and deployment of AI,” Dr Soon said, citing bias exclusion, disinformation, and threats to even elections and democracy as some examples.

Singapore’s Adaptive Policy Approach

Delivering a government perspective, Dr Puthucheary elaborated on Singapore’s policy approach, emphasising a focus on enabling reinvention in the face of technological disruption.

He said AI was “not cooked yet” and Singapore needed a framework that combined building expertise, both in government and the general public, formulating a judicious regulatory framework, while investing in enabling infrastructure for use-cases that might not be imagined yet.

“We don't know what we will need in order to be a significant player in whatever the next round of technology is… We have to have in place a certain set of platforms, mechanisms, and tools that will give us the agility and the confidence that we'll be able to pivot in the future.”

The Evolving Landscape of AI Development

Building on the theme of adaptability, OpenAI’s APAC head of policy, Sandy Kunvatanagarn, shared insights from the industry’s perspective. She told the session that the AI pioneer’s signature product, ChatGPT, wasn’t expected to be the game changer it ended up becoming.

“When ChatGPT was launched, the morning that it went out, very honestly, the company thought it wasn't going to be a big deal.”

Reflecting on the unexpected impact of ChatGPT, she connected this experience to Dr Puthucheary’s cooking analogy, suggesting that we are still at the ‘recipe formation stage’ of AI development.

“If people have ideas for what else would make this a tastier dish, a spicier meal, please, I think everybody's thoughts are super important. Everybody has a role to play,” she said, echoing the open flavour of debate on the topic.

“I talk about it as entering the age of intelligence. It's the age where there's going to be all kinds of breakthroughs, things that we never thought were possible before.

“And some of those things are already starting to happen here in Singapore. There's lots of doctors who are using AI technology here to help with diagnostics.”

As these advancements unfold, the importance of cross-sector collaboration becomes increasingly evident. The session agreed that cooperation at both local and international levels is crucial.

“As all these models and all these innovations enter the world, we believe the governance models also need to evolve alongside all the models. And they need to be based in the science that is driving AI innovation, because we want the governance model to be durable. In order for them to be durable, they need to match where the science is going,” Ms Kunvatanagarn said.

“Government has an incredibly important role to play, but so does civil society. So do the (general) AI Safety Institutes, so do research labs.”

Safeguarding Electoral Integrity in the Age of AI

Responding to this, Dr Puthucheary said that Singapore is actively working on legislation to safeguard electoral integrity from AI-generated threats and manipulations, particularly deepfakes.

The Elections (Integrity of Online Advertising) (Amendment) Bill, introduced in September 2024, is designed to address the risk of digitally manipulated content during elections. The Bill targets content that depicts candidates saying or doing things they never actually said or did, with a focus on ensuring voters are not misled by hyper-realistic AI-generated media.

The bill’s intent was to shore up public trust, he said.

“The electoral process is not only about the actual act of voting, it's about the information that voters have in order for them to make their choices,” said Dr Puthucheary. “And that is then the surface area for manipulation. And it's ripe for manipulation with AI driven tools,” he added.

The Enduring Importance of Human Judgment

Policymakers themselves are feeling the pinch and there were questions about their role in a future dominated by AI.

During the question-and-answer segment, someone asked if AI could replace policymakers in analysing social issues and creating policy.

Dr Puthucheary said AI was already involved in the policy making process.

“Whether it's on the research, synthesising material, criticising, or correcting material, AI and generative AI and large language models are already being used as part of policy making... But there is a point at which you need a humanistic touch.”

Rethinking Our Perception of AI

Dr Puthucheary suggested rethinking our perception of AI ‘hallucinations’ in the future.

“Today, the way in which AI hallucinates is seen as a problem to be fixed. But at some point, if you can have an AI where the sense of imagination and theory of mind suggests that that is a useful thing. In other words, that they are drawing inferences, which the public then subsequently trusts, they are creating positions and synthesising knowledge in a way that we don't then say derisively is a hallucination, but is a synthesis, which is then useful for us, which speaks to that idea of imagination, a theory of mind,” he said.

His point brought home the bigger picture of the way society needs to be prepared to reinvent to tackle unprecedented opportunities, such as advancements in healthcare diagnostics and protein mapping and challenges like misinformation and electoral manipulation. ]

Overall, the panellists agreed that while AI can augment tasks traditionally performed by policymakers and journalists, the human element ultimately remains important.

The insights from the session were a general call for all stakeholders to engage in open dialogue, continuous learning, and collaborative efforts in the face a future with unlimited opportunities posed by AI.

This article is produced as part of a series in conjunction with Festival of Ideas 2024.