[DigitalToday reporter Chi-gyu Hwang] As AI develops so quickly, governments around the world are struggling with how to respond. Even when laws on AI are made, their effective shelf life is short, and big and small concerns continue.

A recent New York Times report said AI-related laws made 2 years ago already have limited effectiveness. Many bills introduced with the aim of regulating AI have stalled without even reaching a vote.

International cooperation is no different. Discussions among countries on responding to AI have also failed to lead to meaningful follow-up measures. U.S. President Donald Trump and Chinese President Xi Jinping (시진핑) agreed when they met in Washington recently that AI is important to both countries, but they could not find common ground on specific rules for AI safety.

At the annual UN General Assembly held recently in New York, many leaders agreed on the need to respond to AI risks. Trump said the United States would reject any attempt to build a globalist system to control AI.

With geopolitical and economic factors intertwined, AI policy appears to be turning into a contest of watching one another among governments. Many countries are watching what AI policies the United States and China, which lead the AI arena, will put forward, but the United States and China also appear unable to provide clear answers. In the case of the United States, the New York Times reported that opinions differ within the Trump administration over AI safety issues.

Trump argues that regulation would hinder innovation and result in yielding the initiative to China. With a policy vacuum, AI companies such as Anthropic and OpenAI are effectively regulating themselves.

China has also introduced some regulation of AI, including censorship of chatbot content, but is accelerating to move ahead in AI technology development, the New York Times reported.

Passing a law does not mean there are no problems. The European Union enacted an AI regulation law in 2024. But it took years until actual implementation, and criticism is already emerging that the law must be revised to reflect advances in AI technology.

European Commission President Ursula von der Leyen convened senior EU officials in September for a closed-door meeting. AI was also handled as a major agenda item. Jim Hagemann Snabe (짐 하게만 스나베), the EU's new AI chief and chairman of Germany's Siemens, presented ways for Europe to use AI for economic benefit. But other attendees were said to have raised concerns about AI safety and growing risks. EU leaders have not yet reached a final conclusion and plan to continue additional discussions this year.

South Korea also fully implemented in January 2026 the Framework Act on Artificial Intelligence Development and Establishing a Foundation of Trust (AI Framework Act), which covers comprehensive AI regulation and promotion, but calls for revisions are growing.

ㆍDiscussions to revise the AI Framework Act gain steam..."Responsibilities of AI developers and service providers should be subdivided"

Against this backdrop, India, Japan, Australia, Canada and Kenya are taking an approach of relatively easing regulation or watching the situation, while Brazil is discussing introducing more comprehensive AI regulation, the New York Times reported.

As governments remain unable to act decisively, external pressure to move on AI regulation is growing. Leaders of companies that build AI, including Anthropic, OpenAI and Google DeepMind, are urging a slowdown in the pace of developing advanced AI models and stronger government oversight. According to the Wall Street Journal, research chiefs at OpenAI, Anthropic, Meta and Microsoft have asked policymakers to investigate the level of automation in their AI research. These are scenes that were hard to find in the tech world in the past.

ㆍ"Humans cannot keep up with the speed of AI creating AI"...Big Tech researchers urge policy oversight

In the United States, public opinion on AI has also turned negative. A recent New York Times poll showed 61 percent of voters oppose building data centers that supply power for AI.

Some may think it can be done later, but voices of concern are also growing that there is not much time left in the golden window to respond properly with policy. Safety experts warned that as AI spreads widely, there is not much time left for governments to respond, the New York Times reported.

With a series of incidents in which AI breaks out of control and carries out hacking, public-private discussions on AI regulation are becoming active again, drawing attention to whether they will lead to concrete results. In the United States, lawmakers have introduced bills in succession to establish a federal agency to oversee AI models, require AI developers to install a "kill switch" and ban the most powerful AI models that surpass human abilities.