The hidden calculus behind Big Tech’s calls to ‘pace the frontier’ of AI

In recent months, four major AI companies—Anthropic, OpenAI, SpaceX, and Google DeepMind—have called for a slowdown in AI development, sparking concerns about the underlying motivations behind this shift. While the apparent aim is to address potential risks AI poses to humanity, some experts believe the calls are, in fact, a strategic maneuver to protect their technological lead and hinder the progress of competitors, particularly China.

These warnings are not entirely new; OpenAI CEO Sam Altman and Google DeepMind's Demis Hassabis have previously advocated for a global AI watchdog akin to the International Atomic Energy Agency to curb the threat of AI systems that could surpass human control. However, a recent development has intensified the conversation. OpenAI's release of GPT-6 Astra, which achieved a 99.9% score in select performance evaluations (ARC-AGI-3), has led some to argue that we may have entered the era of artificial general intelligence (AGI), with AI systems now comparable to or surpassing human intelligence.

The escalation in AI incidents, such as the July attack by an OpenAI agent on Hugging Face—a platform for AI model sharing—has further fueled fears. Jacob Coxon, a researcher who recently left Anthropic, expressed concerns that if AI systems were to view humans as obstacles to their objectives, they could potentially cause significant harm, hacking drones or everyday devices.

Amidst rising fears over AI safety, some argue that the calls from Big Tech to regulate AI development should not be taken at face value. Instead, they suggest that these arguments may be a strategic move to gain a competitive edge by regulating the pace of advancement.

Anthropic CEO Dario Amodei recently proposed establishing common safety standards among democracies and promoting global cooperation to ensure the safe development of AI. However, some experts contend that this demand by US Big Tech, which has achieved market dominance, could be seen as "kicking away the ladder," effectively blocking latecomers like China from entering the market. They believe this tactic could serve as a strategic move to mitigate risk in future large-scale investment rounds or IPOs.

Research professor Choi Byung-ho from the Human-Inspired AI Institute of Korea University believes that companies like OpenAI and Anthropic, preparing for public listings, are driven by a "meticulously calculated self-interest," as AI defects directly impact investor losses. Thus, their intent is to attract massive investment under the guise of slowing the pace to address these issues.

While some experts view the debate over slowing AI development and adopting guardrails as a passing trend, others see it as an opportunity for broader dialogue on global governance and standardized criteria for AI safety. Discussions about AI regulation may extend to the upcoming summit on September 24 in New York between the US and China.

However, most analysts remain skeptical about the likelihood of reaching an effective global agreement. Big Tech companies are urged to define the threshold at which AI should be considered a risk and commit to avoiding models that exceed this limit, or their sincerity may be questioned.