Singapore launch urges reading to tackle AI misinformation

KPMG and the National Library Board have launched Read to Lead: Building an AI-Ready Mind in Singapore, a year-long initiative for professionals and business leaders.

The programme promotes reading as a workplace skill as AI-generated content becomes more common in daily work. Launched with the involvement of Singapore’s Ministry of Digital Development and Information, it is expected to reach more than 2,000 professionals, managers, executives and technicians, along with business leaders.

New polling by KPMG in Singapore and NLB suggests a gap between workers’ confidence and their verification habits as they face a growing volume of AI-produced material. In a dipstick survey of 1,150 PMETs, only four in 10 said they were confident in their ability to distinguish accurate content from AI-generated misinformation.

At the same time, fewer than half said they would check the original source of a statistic before forming an opinion, even when reading an AI-generated summary in a search engine. The poll examined whether professionals critically evaluate information as AI-generated outputs become more widespread.

The organisers frame the initiative around the idea that reading remains a core way to develop judgement, context and scrutiny in knowledge work. Rather than treating reading solely as a personal habit, the programme presents it as a professional discipline employers should support.

Poll findings

The findings come as companies across sectors decide how far to rely on generative AI and other automated tools in research, drafting and decision-making. While these tools can reduce time spent on repetitive tasks, they also raise concerns about accuracy, bias and the erosion of independent judgement if workers rely too heavily on summaries and machine-generated outputs.

Those concerns were echoed in comments from speakers at the launch. Several highlighted the risk that convenience could displace deeper understanding, particularly when staff consume information through algorithmic feeds or AI tools that simplify complex subjects.

“Knowledge isn’t built by consuming more information. It comes from retaining, reflecting on and applying what you’ve learned in meaningful ways. There is no shortcut to substance. AI can accelerate learning and productivity, but lasting value still comes from building strong foundations of knowledge, judgement and critical thinking,” said Voo Poh Jee, Partner, Audit Innovation, KPMG in Singapore.

Show More

One discussion focused on how professionals should respond to the growing volume of online content whose origins may not be obvious.

“We need to recognise that we are increasingly consuming content that may or may not be AI-generated. Algorithms are designed to show us more of what we already like. So, we need to make a deliberate break out of the echo chamber by reading broadly – across different domains and topics to connect the dots, exercise critical thinking and form perspectives that go beyond what is presented to us by AI,” said Gerry Chng, Partner, Head of Cyber, KPMG in Singapore.

“That means we need to spend time to understand AI’s strengths and limitations, exercise meaningful human oversight to understand the outputs, and apply sound judgement,” Chng said.

Reading and judgement

The emphasis on reading as a workplace skill reflects a broader concern that professionals may become more efficient without becoming better informed. For employers, that raises questions about training, internal controls and how organisations define digital literacy.

Warren Fernandez of the S. Rajaratnam School of International Studies linked the issue to perspective and resilience in uncertain conditions.

“The world is changing before our eyes and it can feel unsettling. But when you understand history, you realise that changes in the global order and the uncertainty it gives rise to are not that unusual-and that perspective gives you the confidence and courage to navigate what’s ahead. The joy of reading is a lifelong advantage. No matter how the world evolves, people who know how to find, process and interpret information will always be better equipped to navigate uncertainty,” said Warren Fernandez, Senior Fellow, Head of National Security Studies Programme, and Editor, RSIS Publications, S. Rajaratnam School of International Studies, Nanyang Technological University.

Another theme was the distinction between tasks that can be delegated to AI and decisions that still require human expertise.

“In theory, if a task is low risk and repetitive, there’s little reason for a human to do it. At the other end of the spectrum, high-risk, non-repetitive decisions are not something we would want to outsource to AI,” said Simon Chesterman, David Marshall Professor of Law & Vice Provost (Educational Innovation), National University of Singapore.

“In practice, however, the real challenge is developing expertise. AI can provide shortcuts to polished outputs, but it still takes genuine expertise to distinguish what’s accurate, what’s flawed, and what really matters,” Chesterman said.

Workplace adoption

The programme includes talks, interactive activities and resources designed to help professionals evaluate information more critically. KPMG and NLB also said they will co-develop a practical toolkit on AI literacy, misinformation and cyber risks as part of the wider effort.

Sng Yan Yuan, Chief Operating Officer of Laimi, echoed that focus on implementation, speaking about the role of managers in overseeing AI use within organisations.

“To successfully lead AI transformation, leaders need to understand the technology well enough to guide their teams and spot the blind spots that their less experienced team members may not see,” said Sng Yan Yuan, Chief Operating Officer, Laimi.

“I only trust AI when my own judgement has been encoded into it. To do that, I follow a four-step framework for every AI system I build. First, feed the right data into the system. Second, define what a good output looks like and evaluate the output. Third, pay attention to the performance of the AI-generated work. Finally, create a feedback loop to continuously refine the system,” Sng said.