AI company Anthropic has released a fresh research and policy report suggesting that artificial general intelligence (AGI) could become a reality as early as 2028. In the paper, titled 2028: Two Scenarios for Global AI Leadership, the company also stresses that the United States should work to stay ahead of China in the race to develop advanced AI technologies.
AI capabilities are progressing at a much faster pace:
According to the report, AI capabilities are progressing at a much faster pace than many expected. Anthropic says future systems may soon be able to perform difficult tasks in areas such as scientific research, engineering, cybersecurity, and advanced problem-solving with skill levels similar to trained human professionals.
In its report, Anthropic paints a future where massive data centres powered by advanced AI systems could function like a “nation of geniuses,” helping speed up breakthroughs in science, software creation, and even the development of newer AI technologies. The company believes that whichever country takes the lead in advanced AI could gain significant economic strength along with greater political and military influence on the global stage.
At the same time, Anthropic cautioned that highly capable AI tools could be misused by authoritarian governments for large-scale surveillance, cyber warfare, and tighter control over citizens.
What is AGI?
Artificial general intelligence, often called AGI, refers to a type of AI system that can perform a wide range of tasks with human-like intelligence. Unlike today’s AI tools that are designed for specific jobs, AGI would be capable of learning, reasoning, problem-solving, and adapting across different fields without needing separate training for each task.
United States currently holds an edge:
The report says the United States currently holds an edge because of its control over advanced semiconductor technology and the computing power needed to train next-generation AI models. However, the company warned that this advantage may not last if gaps in chip export restrictions, overseas computing access, and AI model sharing remain unresolved.
Anthropic also argued that Chinese AI companies have managed to stay competitive by relying on computing resources outside China and through what it called “distillation attacks,” a process where outputs from advanced AI systems are allegedly used to build cheaper models with similar capabilities.
Looking ahead, the company outlined two possible scenarios for 2028. In the first, the United States and allied nations strengthen export regulations, crack down on chip smuggling, and rapidly expand domestic AI adoption.
Anthropic believes this approach could help democratic countries maintain an AI lead of roughly one to two years over Chinese competitors.
In the second scenario outlined by Anthropic, weak enforcement of restrictions and continued access to overseas computing resources could help China remain close to the cutting edge of AI development. The company warned that such a situation may give China greater power to influence how the global AI industry evolves in the coming years.