AGI could arrive by 2028: Anthropic
What's the story
Anthropic has predicted that Artificial General Intelligence (AGI) could be achieved by 2028. The company has also warned the US to tighten chip controls and policies to prevent China from taking the lead in this global AI race. The warning comes as President Donald Trump wraps up his 2-day visit to Beijing, his first since 2017.
AI evolution
Anthropic outlines a potential future with AGI
In its latest paper, 2028: Two Scenarios for Global AI Leadership, Anthropic envisions a world where advanced AI systems could soon be capable of performing complex intellectual tasks in science, cybersecurity, engineering and research. These capabilities would be at par with or even exceed those of human experts. The company describes this future as a "country of geniuses in a data center," where AI systems actively contribute to scientific discovery and software development.
Power shift
AI's impact on global power dynamics
Anthropic warns that whoever controls these advanced AI systems will wield tremendous economic, political and military power. The company emphasizes the need for the US and its allies to stay ahead of authoritarian governments like the Chinese Communist Party (CCP). It warns that AI could soon be powerful enough to suppress citizens on an unprecedented scale or even shift the balance of power between nations.
AI dominance
US currently leads in advanced AI development
Anthropic says that the US is currently leading in advanced AI development, thanks to its dominance in compute, high-end chips and infrastructure needed to train frontier AI models. However, the company warns that this edge could be lost quickly if loopholes around chip exports, overseas data centers and model access aren't fixed. It also notes that political systems building these advanced models will ultimately determine how they are deployed globally.
Compute control
The importance of compute and export controls
Anthropic stresses on the importance of compute in developing frontier AI. The company credits export controls by successive US administrations for helping America maintain its edge over China. However, it claims Chinese AI firms have stayed competitive by exploiting loopholes in these controls, using overseas compute access and "large-scale distillation attacks," where outputs from leading US AI models are seemingly used to recreate similar capabilities at lower cost.
Future predictions
Two scenarios for the future of AI leadership
Anthropic outlines two possible scenarios for 2028. In the first, US and its allies successfully tighten controls on advanced chips, prevent smuggling and accelerate domestic AI adoption. This could let democracies stay ahead with American systems being 12-24 months smarter than Chinese rivals. The second scenario sees China staying close to frontier due to weak enforcement, continued access to overseas compute infrastructure and use of American systems for research/development.