- Anthropic researcher Jacob Coxon resigned in September 2026 over concerns that AI development is on a collision course with humanity.
- Tech CEOs warn AI capability is doubling roughly every four months, far outpacing regulatory oversight.
- Unregulated competition among technology companies — and between the US and China — is described as putting global infrastructure at risk.
- The commentary argues that cooperation between the Trump and Xi administrations may be essential to steering the AI revolution safely.
The resignation of Jacob Coxon, a researcher at Anthropic, over concerns that artificial intelligence was on a collision course with humanity has renewed public attention on a question that has been building for years: who, if anyone, is steering the development of the most powerful technology in human history? The headlines that followed his departure focused on a familiar theme — unregulated competition between technology companies, and between the United States and China, is putting global infrastructure at risk at the very least, and threatening human existence at worst.
The anxiety is not confined to researchers. The chief executives of major technology firms have themselves issued stark warnings about the rate of change, with AI capability said to be doubling roughly every four months. That pace leaves little room for the slow machinery of legislation, international treaty-making, or even internal corporate governance to catch up. When capability compounds faster than oversight, the gap between what systems can do and what societies have agreed they should do widens with every cycle.
Why Bilateral Cooperation Matters
The competitive dynamic between Washington and Beijing sits at the center of the problem. Both governments view leadership in AI as a strategic imperative, and both have strong incentives to press forward rather than pause. Yet the same dynamic makes unilateral restraint nearly impossible: any country that slows down risks ceding advantage to the other. That is the classic structure of an arms race, and it is precisely the structure that makes bilateral or multilateral agreement so valuable — and so difficult.
History offers some grounds for cautious optimism. During the Cold War, the United States and the Soviet Union eventually negotiated arms control agreements despite deep hostility, because both sides recognized that certain risks were shared. AI safety, model transparency, and limits on autonomous weapons could plausibly fall into a similar category. The challenge is that AI development is driven largely by private firms operating across borders, not by state arsenals alone, which complicates verification and enforcement.
The Stakes for Markets and Infrastructure
For investors, the debate matters in concrete ways. The largest technology companies — among them Alphabet, Microsoft, and Nvidia — are both the primary engines of AI progress and the entities most exposed to any future regulatory regime. Grid operators, financial exchanges, and communications networks increasingly depend on AI systems, meaning that failures or adversarial uses could transmit shocks through the broader economy rather than remaining contained within the technology sector.
None of this guarantees catastrophe, and predictions of doom have accompanied every major technological transition. But the combination of rapid capability gains, intense commercial rivalry, and geopolitical competition creates a risk profile that markets and governments have not previously had to manage. Whether the Trump and Xi administrations can find common ground on AI governance may determine whether the coming decade is remembered for remarkable progress or for a crisis that cooperation could have prevented.
Comments are closed.