Ahn Byeong-ok
The author is a professor of international relations at Daegu University.
“Did I make you or did you make me?”
This line comes from Isaac Asimov’s robot stories, in which a seemingly perfect machine challenges its creator. Three years after ChatGPT’s debut, huge investments in artificial intelligence and soaring stock prices have dominated the headlines. But a more fundamental problem has slipped into the background. To what extent should society tolerate AI?
Even the European Union, which enacted the world’s first comprehensive AI law, is currently considering postponing parts of the bill that are due to come into force next August. Sustained pressure from the United States and growing opposition in Europe to what critics call overregulation are prompting a reassessment. Yet, as Asimov’s provocation suggests, regulation remains essential if AI is to serve humanity rather than undermine it. The challenge lies in striking a workable balance between innovation and oversight.
US President Donald J. Trump (center) greets Senator Ted Cruz (left) and US Secretary of Commerce Howard Lutnick (right) at the AI signing ceremony in the Oval Office of the White House in Washington, DC on December 11 (EPA/Yonhap News)
The EU’s AI law has been gradually implemented since August 1st last year. We classify the use of AI into four risk categories, each with different regulatory requirements. The “unacceptable risk” category prohibits AI systems that predict criminal behavior or use emotional recognition in workplaces and schools. These provisions took effect in February.
By contrast, rules governing “high-risk” applications, which were originally due to apply from August next year, may be postponed. The European Commission is proposing to postpone the implementation of some provisions until December 2027 and others until August 2028.
High-risk systems include AI used in areas that can have a significant impact on health, safety, or fundamental rights. These require human oversight and strict checks against rights violations. The rules cover AI used in employment and workplace management, access to essential public and private services, credit scoring, legal preparation, and assessing the reliability of evidence. The commission has formally submitted amendments, so any delay would require approval by the European Parliament and the Council of the European Union, which represents member state governments.
The EU has long defined its identity around being the first to comprehensively regulate AI. The law aims to improve transparency by requiring disclosure when using AI, protect fundamental rights, and foster innovation based on risk management. This approach contrasts with the United States and China, which prioritize rapid development of AI.
The law applies not only to companies operating within the EU, but also to companies outside the EU that provide AI services to European users or sell AI-related products in Europe. This breadth of scope explains why jurisdictions such as Japan, Brazil, and the US state of California have adopted rules requiring disclosure of the use of AI. EU regulations have significant global ramifications.
But while the EU seeks to set global standards, the Trump administration has been pressuring Brussels from the beginning to loosen its regulatory stance. In its national security strategy released earlier this month, the US government labeled the EU a “regulatory choke point” and accused EU institutions of violating its sovereignty. The committee harshly criticized companies such as Meta and Company X for imposing large fines on them for violating antitrust laws and for data transparency. Immediately after taking office, President Trump singled out the European Union, which has a trade surplus of more than $200 billion with the United States, calling it an organization “designed to harm America.”
France and Germany, along with major European companies, have echoed the US’s request for a delay. They argue that compliance requires significant staffing and costs, that the law goes beyond what is necessary, and that Europe is at a disadvantage in the AI race with the United States and China.
Responsibility for this decision currently rests with the European Parliament, and opinions are divided. Center-left parties such as the Social Democrats and the Greens oppose the postponement, arguing that the EU has a leading role in establishing global AI governance and should not relinquish its agenda-setting position. Center-right groups such as the European People’s Party support the delay, citing the lack of clear technical standards and guidelines. The EU has already eased some of its green policies in a bid to break free from low growth, and momentum is shifting towards deregulation, raising the possibility of a partial postponement.
Meanwhile, the United States is grappling with a domestic dispute over AI regulation. On December 11, President Trump issued an executive order aimed at advancing AI while preventing states from adopting conflicting rules. This gives the federal government the power to review and challenge state-level AI laws deemed inconsistent with federal policy. The White House opted for executive action after a similar bill stalled in Congress in July due to opposition from some Republicans.
The European Union flag flies outside the European Commission headquarters in Brussels, Belgium, April 9 (Reuters/Yonhap News)
The move sparked a backlash. Gavin Newsom, often mentioned as a future Democratic presidential candidate, said he would sue the order, arguing it violates the state’s authority. Even among President Trump’s “Make America Great Again” supporters, many support AI regulations because they are concerned that job losses due to automation will outweigh job creation.
Nevertheless, Trump continues to champion AI development alongside Big Tech companies, framing the issue as an “AI arms race” with China that the United States must win. The US government has agreed to cap tariffs on EU goods at 15%, but tariffs on steel and aluminum remain at 50%. The US has repeatedly called on the EU to ease green and AI regulations as a condition of easing trade pressures.
The EU has long been accustomed to exercising regulatory powers and is unlikely to abandon its framework solely due to US pressure. Regulatory tensions between Washington and Brussels are therefore likely to continue.
This article was originally written in Korean and translated by a bilingual reporter using generative AI tools. It was then edited by a native English speaking editor. All AI-assisted translations are reviewed and refined by our newsroom.
