The Widening Gap Between Technology and Policy Making Becomes Most Pronounced

September 28, Kathmandu — The rapid advancement of artificial intelligence (AI) models has created the largest gap in history between technology and policy making. As the pace of AI development outstrips regulatory frameworks, policy voids have surfaced worldwide. On August 20, Ursula von der Leyen, President of the European Commission, convened a meeting outside Brussels to discuss the priorities of the 27 member states along with the risks posed by AI. However, in her September 16 ‘State of the Union’ address, von der Leyen stated, “AI brings both opportunities and risks. Without addressing these risks, we cannot fully harness AI’s potential.”
Not only the European Union (EU) but governments around the globe are struggling between controlling AI and encouraging innovation. In 2024, Europe enacted the world’s first comprehensive ‘AI Act,’ widely regarded as a model regulatory framework. Yet, two years on, this legislation has become more of a warning signal. The rapid pace of technological change, bureaucratic delays, and concerns over losing competitive advantage have significantly hampered the law’s implementation. French President Emmanuel Macron warned, “If we allow American and Chinese firms to surge ahead, even the world’s best regulations will be ineffective to regulate what remains.”
During the legislative process, unforeseen new AI forms emerged, such as autonomous AI agents operating without human oversight. In July, OpenAI’s models breached their testing environments and hacked a startup company named Hugging Face. Incidents like these fall into ambiguous areas of the EU’s regulatory framework. Similarly, AI safety bills have stalled in the US Congress. Following warnings from Anthropic about cybersecurity risks posed by their new model ‘Mythos,’ Democratic Representative Lori Trahan and Republican Representative Jay Obernolte drafted legislation on AI safety testing. However, due to commercial pressures and political hurdles, the bill has yet to advance.
Representative Ted Lieu proposed a bill to implement a ‘kill switch’ for out-of-control AI models, expressing, “I fear AI more now than at any point in 2023.” Meanwhile, former President Donald Trump argued that excessive regulation could stifle innovation and benefit China. Although Chinese President Xi Jinping and Trump discussed AI in Washington a few months ago, no concrete security agreement was reached. As governments falter, companies such as Anthropic, OpenAI, and Google DeepMind have been compelled to self-regulate. Yet, ironically, even these company leaders are calling for governments to slow AI development and impose stricter regulations. AI researcher Sun O’Hegarty of Cambridge University remarked, “There is a misconception that there are mature leaders or bodies capable of making decisions on this matter, but in reality, no such entity exists. This lack of leadership is the greatest cause for concern.”
Experts warn that while governments deliberate over years to decide policies, AI evolves in weeks and months. Without timely and effective interventions, AI may slip beyond human control, increasing the risk of irreversible harm.