Concerns over the rapid development of artificial intelligence (AI) have intensified recently, with experts warning about potential risks while some leaders call for caution against overregulation. Last week, Jacob Coxon, a researcher at AI company Anthropic, resigned and issued urgent warnings on social media, saying that leading AI developers are "gambling with our lives" and suggesting that AI could pose an existential threat within the decade.

Dario Amodei, Anthropic's CEO, followed with a detailed open letter expressing fears that autonomous AI systems capable of self-improvement could soon dominate the internet, potentially causing hundreds of billions of dollars in damage. Amodei urged AI companies to slow down their development processes to better align and safeguard AI models. He also called for international cooperation, particularly between the United States and China, to address acute risks associated with AI advancement. The letter gained endorsements from prominent figures in the AI field, including Elon Musk and OpenAI CEO Sam Altman.

In response, former U.S. President Donald Trump dismissed these safety concerns as a “hoax” and argued against stringent AI regulation, warning that such restrictions could allow China to surpass the U.S. in AI innovation. Nvidia CEO Jensen Huang also criticized Coxon's warnings, calling them "not grounded in science." Their position reflects a broader debate on whether AI should be primarily viewed as a disruptive technology that fosters innovation or as a potentially destructive force.

Historically, technological breakthroughs like the steam engine, the airplane, and personal computers have disrupted economies but ultimately contributed to job creation and economic growth by increasing productivity. A recent study from Stanford’s Institute for Economic Policy Research found no evidence that AI adoption has led to reduced job postings; on the contrary, employment grew by 10 percent among firms adopting enterprise AI over two years.

However, some experts argue that framing AI merely as a disruptive technology might underestimate its unique and profound risks. Comparisons have been made to groundbreaking yet dangerous scientific advances such as nuclear research or controversial gain-of-function studies on viruses. Past U.S. presidents have grappled with similar dilemmas: Barack Obama implemented a pause on certain risky virus research in 2014 to understand its dangers better, while Harry Truman established the Atomic Energy Commission in 1946 to regulate nuclear technology and promote civilian use.

Both historical analogies offer lessons but also have limitations. For example, although Obama’s moratorium slowed research in the U.S., some activities shifted overseas, potentially under less secure conditions. Similarly, while the Atomic Energy Commission helped control nuclear technology in its early years, nuclear power remained contained in a controlled environment, unlike AI, which has been widely accessible from its inception.

Given these challenges, some experts argue that slowing AI development is unrealistic. Instead, they suggest accelerating efforts to understand and manage AI risks, developing strategies to harness AI safely and mitigate threats it might pose. Although previous scientific advances have elicited fears about humanity’s destruction, past experiences do not guarantee immunity from future dangers.

The AI debate remains complex, balancing innovation, economic opportunity, and safety concerns, with no consensus yet on the optimal regulatory approach. The evolving discourse emphasizes the critical need for careful policy decisions in a technology landscape that is advancing rapidly and globally.