What happened
Industry experts and Singaporean officials are proposing that the AI sector adopt international safety frameworks similar to those used in civil aviation and nuclear non-proliferation. On September 17, Singapore’s Minister for Digital Development and Information, Josephine Teo, suggested that AI may require a global safety regime akin to the International Civil Aviation Organization (ICAO) to foster trust and standardize accident investigation and safety protocols. Experts cited in the report argue that such a framework could accelerate the development of independently verifiable safety assurances, moving beyond the current fragmented approach to .
The proposal, highlighted by The Straits Times, suggests that the AI industry could compress the timeline for safety regulation by adopting established international models. Jonathan Zhang, president of the Singapore AI Association, noted that the industry can bypass the decades-long evolution of aviation rules by leveraging existing knowledge of what 'internationally trusted' safety looks like.
Hassan Shahidi, CEO of the Flight Safety Foundation, emphasized that aviation safety is built on the principle of determining the cause of accidents to prevent recurrence, rather than assigning blame. He urged the AI industry to build safety architecture now rather than waiting for a catastrophe.
Singaporean officials and industry leaders, including Dennis Yap of AI Seer, suggested that a tiered regulatory approach—similar to financial sector oversight—could be effective. Under this model, large-scale AI developers would face the most stringent obligations, such as mandatory independent testing and incident reporting, while smaller entities would face lighter requirements to avoid stifling innovation.
The report notes that while leaders like Dario Amodei (Anthropic) and Sam Altman (OpenAI) have called for development slowdowns, these calls face significant geopolitical opposition. President Trump and Chinese officials have expressed concerns that such measures could disadvantage their respective nations in the global AI race.
Source details: straitstimes.com ↗
Why it matters
Establishing a global safety framework is critical because the current lack of consensus on AI risk management—particularly between the U.S. and China—threatens to stall progress or lead to catastrophic failures. By applying lessons from aviation, where safety is embedded in design and operation, and nuclear non-proliferation, which relies on bilateral and multilateral treaties, the industry aims to move from reactive measures to proactive, standardized safety architectures. This shift is essential for public trust and for creating a predictable environment where innovation can proceed without waiting for a major disaster to trigger regulation.
The primary challenge identified is the lack of global consensus on the priority of safety versus speed. Without a unified approach, the industry risks a 'race to the bottom' where safety is sacrificed for competitive advantage.
The aviation model provides a blueprint for international cooperation, specifically through the ICAO framework, which allows for standardized accident investigation and safety protocols across borders. Applying this to AI could provide a neutral, verifiable way to manage risks that transcend national boundaries.
The nuclear non-proliferation model, specifically the Strategic Arms Limitation Talks, is cited as a necessary precedent for U.S.-China engagement. Experts argue that without the commitment of these two superpowers, any global initiative is unlikely to succeed.
Practical implementation of these ideas is already underway in Singapore, which has established the AI Verify toolkit and the Singapore Institute. These initiatives serve as a testbed for the type of technical audits that could eventually become part of a global regulatory standard.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
Which of these is a common misconception about AI Ethics?
What to watch next
The upcoming meeting between U.S. President Donald Trump and Chinese President Xi Jinping on September 24 is a key event, as is expected to be on the agenda. Observers are looking for signs of whether the two nations can move past competitive rhetoric to establish shared risk-management protocols. Additionally, the evolution of Singapore’s AI Verify toolkit and the Singapore AI Safety Institute will be monitored as potential models for how neutral, third-party testing infrastructure can support future international enforcement of AI standards.
The September 24 meeting between President Trump and President Xi Jinping is the most immediate indicator of whether the U.S. and China will pursue formal, bilateral discussions.
Future developments regarding a potential UN-sponsored treaty or agency for AI, modeled after the ICAO, will be a significant indicator of long-term global policy direction.
The ongoing work of the International Scientific Exchange on , which has been active since 2025, remains a key forum for tracking how nations are aligning on these safety mechanisms.
The industry will be watching to see if large AI developers voluntarily adopt independent testing and reporting standards, which could serve as a de facto industry standard in the absence of formal government regulation.