The real test starts when AI affects money
For many firms, the first phase of AI adoption was “low stakes”, which was limited to pilots and productivity tools used mainly by technical teams. But that phase is over. Companies beyond regulated industries are adopting AI in core business functions that directly affect customers, such as the interface on your banking app and the security on your office computer.
IMDA’s 2025 pulse survey found that nearly three in four workers in Singapore already use AI tools at work and 85% of those users say that AI helps them work faster and better. But as AI evolves and autonomous agents increasingly enter critical workflows, the stakes of AI governance are escalating. It is one thing to use AI to work better and faster; it is another to trust it to perform critical business functions, like retrieving sensitive information, with relative autonomy. Writing faster is not the same as deciding better. This is the defining feature of the next phase of AI development, in which AI tools are increasingly deployed in business-critical situations.
If Singapore wants AI to lift productivity across the economy, organisations will have to trust it enough to use it in everyday operations, particularly as we enter a new and exciting phase of AI maturation. But research shows that data protection and security issues remain prevalent with AI, undermining trust and creating massive liabilities. The PDPC, for example, reported a 41% increase in large-scale data breaches in Singapore in 2024, while AvePoint’s global research found that 75% of organisations experienced a data breach in the same period.
AI has a data protection problem, which has led to an even more fundamental trust problem. Organisations and leaders across Singapore will have to address these issues and rebuild trust to get the most out of AI technology.
See also: US-backed Philippines AI hub could draw at least US$10 bil in investments, says official
Autonomy without ownership is an economic risk
Singapore’s Model AI Governance Framework for Agentic AI makes it explicit that organisations deploying agentic AI systems need meaningful human accountability and controls to enable adoption with greater confidence.
This has a simple implication for business leaders and managers. Every material AI system needs a human owner, a defined scope and clear visibility on permissions. Someone must decide what data agentic AI has access to, what actions the system is permitted to support, and when a person must approve or override the result. Otherwise, trust is eroded and risks surface in familiar ways: sensitive data being distributed, payments going out that should have been stopped, customer records being altered without sufficient review, or staff assuming the system has handled something when no one is accountable for the outcome.
Still, global research shows that organisations are struggling to implement this kind of AI governance at scale. A new study by Omdia finds, for example, that governance and compliance are the biggest barriers to AI adoption today.
See also: Amazon inks US$17.5 bil loan as AI borrowing frenzy grows
True trust in AI is only possible when an organisation can do three things simultaneously: precisely control what AI can access, govern and audit every action that AI takes, and recover instantly when something goes wrong. By maintaining this continuous oversight, the trust layer transforms AI from a source of potential economic risk into a durable business accelerator. When we bridge the gap between autonomy and accountability, we move beyond speculation and into the era of scaled, secure enterprise AI.
Trusted deployment can be Singapore’s edge
Just as Singapore is a hub for finance and logistics, it need not outspend the US or China on frontier models. Its stronger and more realistic ambition is to become the place where AI can be deployed with confidence across sectors, in banks, factories, hospitals, government services and SMEs alike. In other words, what will keep Singapore ahead is the effective diffusion of AI across our best industries, enabling Singapore businesses to retain their world-class competitiveness.
That confidence rests on clearer rules, traceable accountability, high-quality data and a workforce trained to use it. The Singapore Budget 2026 leans exactly into this agenda, pairing AI research funding with support for enterprise adoption, worker upskilling and tight national coordination.
Trust is what turns AI spending into AI adoption, and adoption into economic value. Singapore is already an early mover. The most important task now is to ensure the ecosystem moves with sufficient discipline to be trusted at scale.
Govern first, code second. That is how Singapore turns ambition into advantage.
Tianyi Jiang is the CEO of AvePoint
