Abstract:
Jacob Coxon, a former artificial intelligence researcher who has worked for OpenAI and Anthropic, recently issued a strong warning against the global artificial intelligence competition. In his latest interview, he said that he is "90% sure" that China has placed intelligence personnel within OpenAI and Anthropic, and believes that as competition in super-intelligence research and development continues to escalate, it is almost impossible for intelligence agencies in various countries to ignore these companies that master the most cutting-edge technologies.
Coxon said that in his opinion, the world's top artificial intelligence laboratories have become an important target of geopolitical competition. Since advanced AI technology may have a huge impact on the economy, military, cybersecurity and scientific research fields in the future, governments around the world have strong incentives to obtain relevant information. He believes it is unrealistic to assume that these institutions are not penetrated at all.
In addition to expressing concerns about security issues, Coxon also disclosed some cultural details within Anthropic. He revealed that many Anthropic employees would call themselves “Ants.” This name comes from the abbreviation and derivative usage of the company name Anthropic, and is also regarded by some employees as an internal identity to jointly pursue the goal of superintelligence.
Coxson has previously received widespread attention for publicly criticizing the direction of the AI industry. As a researcher who has worked at OpenAI and Anthropic for many years, he has long been involved in research related to advanced artificial intelligence systems. After leaving Anthropic, he published a series of public articles, accusing the industry of promoting super-intelligence research and development at an excessive speed, while safety issues have not received attention commensurate with technological progress.
He believes that currently leading AI companies are locked in fierce competition, hoping to be the first to build super-intelligent systems with self-improvement capabilities. In order to compete for technological leadership, companies continue to increase the speed of research and development, but the industry has yet to find a mature solution for how to ensure that future systems are always within human control.
Coxon said that some researchers at the core of the industry sincerely believe that the emergence of superhuman intelligent systems in the next few years is not a distant fantasy, but a real development path. He warned that people generally underestimate the speed of development and potential impact of related technologies. Once systems with a high degree of autonomy gain greater capabilities, they may have the ability to invade computer systems, control critical infrastructure, and gain access to real-world resources and influence.
Discussions surrounding the risks of superintelligence have continued to heat up recently. Many researchers and managers from OpenAI, Anthropic and other top research institutions have publicly discussed the existential risks that artificial intelligence may bring. Some industry insiders advocate establishing a stricter safety regulatory framework and slowing down the development of cutting-edge models to address controllability issues first.
However, some people in the industry believe that over-emphasis on extreme risks may ignore the huge potential brought by artificial intelligence in fields such as scientific research, medical care, education and productivity improvement. They believe that strengthening safety measures and continuing to promote technological innovation are not inconsistent.
As companies such as OpenAI, Anthropic, Google DeepMind, and xAI continue to invest huge sums of money in building next-generation models, debates about artificial intelligence safety, national competition, and the future direction of superintelligence continue to escalate. Coxon’s remarks about intelligence penetration and superintelligence competition have once again pushed the global artificial intelligence industry into the focus of public opinion.
Comments