Artificial intelligence is moving from experimentation to deployment across Asia at a remarkable pace. Companies are rolling out AI agents to handle customer support, improve productivity and manage risk. Universities are launching new programmes in artificial intelligence and data science.
Consumers, meanwhile, are becoming increasingly comfortable interacting with AI-powered systems in their daily lives. Yet amid the rush toward automation, Professor Eric Paringit of the University of the Philippines believes the most important question is not how powerful AI can become, but whether it remains grounded in human values.
Speaking at AIBC Asia in Manila, Paringit reflected on how dramatically the relationship between technology and decision-making has changed over the past decade. For much of his research career, obtaining reliable data was often the biggest challenge. Today, the challenge has shifted.
“During my time as a researcher, data availability was an issue. Now that is not the case anymore,” he said. “The challenge today is coping with the right methods, the right approaches and the right decisions.”
The shift is visible across sectors. From disaster management and governance to finance and business operations, organisations now have access to unprecedented amounts of information generated through sensors, satellites and digital platforms. AI, Paringit argues, arrived at a moment when tools capable of processing and interpreting that information became increasingly necessary.
Universities are adapting to that reality. Institutions, including the University of the Philippines, have introduced programmes in data science, analytics, and artificial intelligence to prepare students for a world where data-driven decision-making is becoming standard practice. At the same time, businesses across Asia are embracing AI at a speed few predicted just a few years ago.
A survey commissioned by UiPath shows that 42 per cent of organisations in Southeast Asia have already deployed AI agents, while another 44 per cent plan to implement them within the next 12 months. The technology is most frequently used for customer support automation, fraud and risk management, and productivity enhancement.
The benefits appear tangible. Across Southeast Asia, many organisations say the technology is helping them make better decisions and improve productivity. At the same time, consumers are becoming more comfortable interacting with AI, whether that’s through customer service tools or digital assistants acting on their behalf.
But acceptance does not automatically mean trust. Even among users who are open to AI, there is still a strong preference for transparency. People want to know when they are interacting with an AI system, and many still expect the option to speak with a human when needed. And for Paringit, this is also a reminder that the conversation around AI should not focus solely on what the technology can do, but on how it should be used.
“There had been a great extent of philosophical discussion on the influence of artificial intelligence, including the use of agentic AI in how we manage our affairs as humans,” he said. “There are questions about how influential it could be, how enabling it could be, but at the same time how meddling it could be as well.”
Those questions become increasingly important as AI agents move beyond routine tasks and begin participating in processes that influence decisions, behaviour and human interactions. Rather than slowing innovation, however, Paringit believes the focus should be on responsible deployment.
“Research and development is not only about coming up with tools,” he said. “It is also about trying to test them before they could be deployed on a massive scale, so that we will be able to have a calculated way of assessing whatever benefits or even threats they may have.”
That emphasis on testing and evaluation reflects his belief that human judgment must remain central, even as AI systems become more capable. The goal should not be to replace people, but to build systems that support them.
“What I see is this: humans should document as much as we can the values and virtues that we have in order to reflect those values and virtues in the AI and AI agents that we develop,” he said. For Paringit, the future challenge is not merely creating smarter systems. It ensures that those systems mirror the principles society wants to preserve.
“That way, whatever we entrust to AI and agentic AI, it will capture the way we behave, cope and understand as humans,” he said. “It will not be as different from the very values that we espouse.”
