Anthropic warns human-level AI by 2028, urges US export crackdown on China
By 2028, AI could match expert humans in science and cybersecurity, Anthropic says.
Anthropic, the creator of Claude AI, published a research paper titled “2028: Two Scenarios for Global AI Leadership,” warning that machines with human-level intelligence could arrive by 2028. The company calls on Washington to tighten export controls on China to prevent “distillation” of AI models and close gaps in semiconductor restrictions. In its optimistic scenario, the US and allies maintain a 12-24 month lead by enforcing chip smuggling bans and accelerating domestic AI adoption, enabling democratic nations to lead in science, engineering, and cybersecurity. In the pessimistic scenario, weak enforcement allows China to match US capabilities, potentially enabling mass surveillance and digital repression.
Critics, including Stanford fellow Alvin Wang Graylin, argue that Anthropic’s arms-race framing is self-interested and counterproductive, especially as Presidents Trump and Xi concluded technology talks in Beijing that included discussions on AI guardrails and Nvidia H200 chip shipments. While Anthropic stresses the economic, political, and military advantages of AI leadership, the timing of its report (released during the summit) highlights the tension between industry warnings and diplomatic cooperation. The company’s recommendations aim to preserve America’s current edge in high-end chips and compute power, but experts warn that conflict-focused narratives could derail safety collaboration.
- Anthropic claims human-level AI could arrive by 2028, capable of expert-level work in science, engineering, and cybersecurity.
- The paper outlines two 2028 scenarios: US leads with 12-24 month advantage if export controls are tightened, or China catches up if enforcement is weak.
- Critics call the arms-race framing irresponsible, especially as Trump and Xi discussed AI cooperation and Nvidia H200 chip shipments in Beijing.
Why It Matters
Anthropic’s warning highlights the growing AI race between US and China, with potential geopolitical and ethical consequences.