SINGAPORE — Singapore has warned that safeguards introduced by artificial intelligence companies, while important, may not be enough to manage the growing risks posed by increasingly powerful AI systems.
Minister for Digital Development and Information Josephine Teo said companies have introduced measures such as content filters, user verification and pre-release testing. However, she argued that these protections can be bypassed, particularly when people turn to openly available AI models.
‘No Single Safeguard Will Be Enough’
Teo said Singapore needs a multi-layered approach to AI safety rather than relying solely on safeguards built into individual AI products.
That approach, she said, should include stronger cybersecurity, clear limits on what AI systems are permitted to do, appropriate human oversight and the ability to continuously test and monitor AI as the technology evolves.
The minister stressed that governments and technology companies need to better understand what increasingly capable AI systems can actually do before deciding what additional restrictions or safeguards are necessary.
AI Risks Are Moving Beyond the Technology Companies
Singapore’s position reflects growing concern that AI safety cannot be left entirely to the companies developing the technology.
As AI systems become more capable and autonomous, potential risks can extend beyond individual applications and into cybersecurity, fraud, misinformation and other areas where misuse could have wider consequences.
Teo has previously compared the need for AI safeguards to aviation safety, arguing that multiple layers of standards, testing and oversight are necessary to build public confidence in powerful technologies.
Singapore Pushes International Cooperation
Singapore also believes AI safety requires cooperation across borders.
Teo has said no country can address AI safety on its own because the most advanced systems are developed and deployed internationally. Singapore has therefore been working with governments, researchers, technology companies and AI safety organisations to better understand emerging risks.
The country’s approach includes initiatives such as AI Verify, a testing toolkit designed to help organisations evaluate aspects of AI systems, as well as the Singapore AI Safety Institute, which focuses on testing and evaluating advanced AI models.
The Bigger Question: Who Should Be Responsible for AI Safety?
The debate is increasingly moving beyond whether AI companies should build safer systems to the broader question of how governments, businesses and independent experts should share responsibility for managing AI risks.
Singapore’s latest position is clear: company safeguards remain important, but they should form only one layer of a much wider safety framework.
As AI becomes more powerful, policymakers argue that safeguards must evolve just as quickly.
OnPointGist will continue to follow developments in AI regulation, safety and the global debate over responsible artificial intelligence.
