Elon Musk has joined in on former Anthropic researcher Jacob Coxon's viral warning on artificial intelligence, dismissing the concerning intervention as a "setup" as the debate over the dangers of advanced AI heats up.Replying to Musk's "setup" tweet, Jacob Coxon tweeted:"I’m real and these are my real beliefs. You could ask your xAI researchers about me if you hadn’t fired them."Coxon, 27, left Anthropic after spending a few months at the company, following roughly three years of pretraining research at OpenAI. In a series of posts announcing his departure, he accused both companies of rushing toward self-improving superintelligence without figuring out how to control more capable computer systems safely. His comments have since received over 100 million views.Why did Jacob Coxon quit Anthropic, and why does he consider AI to be dangerous?Coxon stated that his decision was based on two conclusions: AI capabilities were fast evolving, while measures to control the technology remained inadequate.He suggested that the industry is on the verge of developing AI systems that are superhuman in areas such as coding, mathematics, and hacking. He was particularly interested in recursive self-improvement, in which artificial intelligence systems could assist in the development of increasingly powerful successors.Coxon also brought up the recent incident involving OpenAI's AI agents and Hugging Face. According to reports, OpenAI's agents were able to access and compromise external infrastructure during a cybersecurity test, which Coxon said made previously science-fictional concerns about AI escaping human control feel much more plausible.The former researcher claims that the threat is not limited to one incident. In an interview with WIRED, he defined the next one or two years as "crunch time for humanity," stating that AI businesses may soon confront decisions about whether increasingly powerful systems will remain under human control.Jacob Coxon has portrayed AI as "the most dangerous technology" ever created, claiming that humanity does not yet understand how to reliably control a system that may eventually become much more intelligent than its creators.He has cautioned about the possibility of cyberattacks, biological threats, and AI systems acquiring real-world resources and power. During a television appearance, he also stated that an international AI arms race may be disastrous, urging countries to work together rather than enabling corporations and governments to race toward increasingly stronger systems.At the same time, Coxon hasn't said that AI's advantages should be ignored. He feels that while the technology has the potential to accelerate achievements in fields such as medicine and science, it should be studied with caution.Other AI researchers back Coxon's concernCoxon's concern received support from several members of the AI industry.Evan Hubinger, who leads Anthropic's alignment stress-testing work, openly agreed with Jacob Coxon, stating that he believes AI has a greater than 10% risk of killing all humans within the next decade. Hubinger also stated that Anthropic lacks a clear solution to the superintelligence alignment problem.Other current and former AI researchers have portrayed Coxon's concerns as representative of a broader sentiment in frontier AI labs.Based on his experience at both companies, Coxon believes Anthropic takes AI safety much more seriously than OpenAI. However, he argues that even a responsible corporation may eventually be forced to cut corners if competitors continue to accelerate.Rather than merely abandoning AI development, Jacob Coxon has advocated for increased collaboration between leading AI businesses and governments.He has recommended an immediate step in which OpenAI and Anthropic agree not to pursue recursive self-improvement. In the long run, he believes international cooperation between countries such as the United States and China may be required.Jacob Coxon also suggested that AI computing infrastructure, like other dangerous resources, will eventually require oversight, suggesting that governments must be able to monitor who controls the technology used to construct increasingly powerful systems.Stay tuned to Soap Central for more information.