SINGAPORE — Concerns over the artificial intelligence boom are becoming broader and more serious, with investors and policymakers in Singapore warning that the risks now extend well beyond inflated technology valuations.
At two major conferences in Singapore on Wednesday (Oct. 7), discussions about the AI industry included fears of a potential market bubble, disappointing returns on enormous investments and the possibility that increasingly autonomous AI systems could become difficult for humans to control.
The debate comes as AI-related companies continue to drive much of the strength in global stock markets, even as higher energy prices, elevated interest rates and instability in bond markets add pressure to the wider economy.
Singapore minister warns of loss-of-control risks
Singapore Foreign Minister Vivian Balakrishnan said the risks surrounding autonomous AI systems should not be underestimated.
Speaking at the Milken Institute Asia Summit, Balakrishnan warned that claims of having complete control over autonomous AI agents should be treated with scepticism.
He said the potential loss of control over increasingly capable systems was a genuine threat, particularly as AI agents become more autonomous and capable of carrying out tasks with less human intervention.
Balakrishnan also warned that AI could be misused by both governments and non-state actors to develop new forms of weaponry.
The risks, he said, could involve chemistry, biology and weapons of mass destruction.
Singapore has been pushing for international rules governing AI safety, arguing that technological development needs to be accompanied by safeguards strong enough to manage the risks created by increasingly powerful systems.
Bioterrorism emerges as another concern
The potential misuse of AI in biology was also highlighted at the Forbes Global CEO Conference in Singapore.
Kiran Mazumdar-Shaw, founder and chairperson of Indian biotechnology company Biocon, said AI could be used for beneficial scientific research but could also give malicious actors new capabilities.
She warned that bad actors could potentially use AI to help design viruses or other pathogens, describing bioterrorism as a major existential concern.
Mazumdar-Shaw argued that stronger guardrails and governance were needed, saying she did not believe the industry could rely on self-regulation alone.
The warnings reflect a growing debate internationally over whether AI safety measures can keep pace with rapidly improving models and autonomous agents.
Last month, leaders of several major US AI laboratories reportedly called for greater caution over the development of increasingly capable systems, warning about the possibility of AI improving rapidly and moving beyond effective human control.
Investors are increasingly questioning the AI trade
Alongside safety concerns, financial risks remain a major source of uncertainty.
The AI boom has become one of the biggest drivers of investor enthusiasm and stock-market gains, with a relatively small group of major technology companies accounting for a significant share of market performance.
But the enormous sums being spent on AI infrastructure have raised questions about whether companies will eventually generate enough revenue and productivity gains to justify the investment.
Rohit Sipahimalani, chief investment officer of Singapore state investor Temasek, said the AI investment narrative could become a problem for markets if safety concerns, regulation or weak returns cause businesses to reconsider their spending.
He pointed specifically to the possibility that companies may fail to see sufficient return on investment (ROI) as they move further into 2027.
“The biggest risk to the market is the AI narrative,” Sipahimalani said, while acknowledging that the narrative remained strong.
The concern is not necessarily that AI technology itself will fail. Rather, investors are questioning whether the market’s expectations have moved ahead of what companies can realistically deliver.
Ray Dalio calls AI a ‘classic bubble’
Bridgewater Associates founder Ray Dalio offered an even more cautious assessment.
Dalio described the current AI boom as a “classic bubble”, arguing that investors need to distinguish between how transformative a technology may be and whether it represents a good investment at current valuations.
He also pointed to rising interest rates as a potential trigger for a correction.
Higher borrowing costs can make speculative investments less attractive and put pressure on companies whose valuations depend heavily on expectations of rapid future growth.
Dalio said the effects of higher rates had not yet fully damaged other parts of the market, but warned that the pressure could intensify.
Once higher rates begin to have a stronger effect, he said, the AI bubble could start to deflate.
Singapore has a lot at stake
The debate carries particular significance for Singapore because the city-state is positioning itself as a major hub for AI, data centres and digital infrastructure.
The government has been encouraging investment in advanced computing while simultaneously trying to manage constraints involving electricity, water and land.
Singapore’s digital economy was worth S$144.1 billion in 2025, equivalent to 19.3 per cent of GDP, according to the Ministry of Digital Development and Information.
The rapid expansion of AI is also increasing demand for computing capacity, placing additional pressure on the country’s limited resources.
Senior Minister of State for Digital Development and Information Tan Kiat How said Singapore’s goal was not simply to maximise computing capacity, but to obtain the greatest value from the computing resources available.
The government has therefore been introducing rules covering major data centres and cloud-service providers, with a focus on security, resilience and sustainability.
AI is both an opportunity and a vulnerability
Balakrishnan acknowledged the contradiction facing Singapore and other economies.
AI has helped support economic growth and productivity at a time when globalisation is being disrupted by geopolitical tensions, wars and supply-chain problems.
But the same technology could create new economic and security vulnerabilities if expectations become excessive or if AI systems are deployed without adequate safeguards.
He compared the current technological transformation with earlier periods of rapid industrial development, when innovations such as railroads, electricity and steel created enormous economic opportunities while also producing speculative bubbles.
His warning was straightforward: technological progress does not guarantee that every investment associated with it will succeed.
The debate is moving beyond ‘Will AI succeed?’
The latest discussions suggest that the question surrounding AI is becoming more complicated.
It is no longer simply about whether artificial intelligence will transform industries.
Policymakers are increasingly asking how quickly it should develop, who should control it and what happens if its capabilities outpace existing safeguards.
Investors, meanwhile, are asking whether the enormous amounts of money flowing into AI infrastructure will eventually generate sufficient returns.
Those two concerns could become increasingly connected.
If governments impose stricter safety rules, development costs could rise or certain applications could be restricted. If companies fail to see expected productivity gains, spending could slow. And if financial markets begin to doubt the growth story, the effects could spread well beyond the technology sector.
For Singapore, which is simultaneously embracing AI and trying to build safeguards around it, the challenge is finding the right balance.
The message emerging from this week’s forums was not that the AI revolution is ending.
Instead, it was that the bigger the AI boom becomes, the more carefully its financial, social and security risks will have to be managed.
