In Brief:

• Former Ripple executive Emi Yoshikawa highlighted Jacob Coxon’s resignation and his allegations concerning safety practices at leading AI laboratories worldwide.
• Coxon claimed Anthropic researchers recognize existential dangers, yet competitive pressure keeps them advancing capable artificial intelligence systems despite those risks.
• His proposals include international regulatory agreements and temporary capability restrictions while governments establish stronger global safety and accountability frameworks collectively.

Former Ripple executive Emi Yoshikawa has highlighted AI safety allegations raised by former Anthropic researcher Jacob Coxon. According to Yoshikawa, Coxon’s resignation statement attracted attention as developers examined his concerns about artificial intelligence systems.

Her response carried Coxon’s allegations beyond AI circles and placed them before fintech investors. Coxon spent three years conducting neural network pretraining research at OpenAI and Anthropic before leaving the latter company. He accused both laboratories of pursuing self-improving superintelligence without adequate protections against catastrophic outcomes.

“I resigned from Anthropic,” Coxon wrote while explaining his concerns about the development strategies followed by both companies. He claimed neither company was acting responsibly while pursuing systems that could eventually improve independently.

Moreover, Coxon described the competition between leading AI laboratories as a dangerous race involving significant risks for humanity. Corporate pressure, he argued, encourages developers to prioritize technological leadership over caution, transparency, and public safety.

Also Read: Ripple Engineer Reveals Mathematical Verification Push for XRP Ledger

Coxon Claims Anthropic Developers Understand the Risks

According to Coxon, researchers developing advanced models privately recognize that artificial intelligence could pose an existential threat to humanity. However, he claimed employees at OpenAI and Anthropic hold different levels of awareness regarding these dangers.

Coxon alleged that many OpenAI employees do not completely understand the possible consequences of creating highly capable artificial intelligence. By comparison, Anthropic employees reportedly recognize the risks but fear losing the competitive race to rival laboratories.

Consequently, he believes informed researchers accelerate development because slowing their work could strengthen another company’s position. His allegations raise concerns about whether internal safety teams influence major corporate development decisions. Additionally, Yoshikawa’s reaction linked the resignation with financial concerns among technology investors and venture capital firms.

However, Coxon’s claims reflect his personal assessment of practices within Anthropic and OpenAI rather than independently confirmed findings. Neither company provided a detailed response within the account Yoshikawa shared.

Coxon encouraged engineers to consider the wider consequences of their research. He also urged regulators to establish international agreements governing advanced AI development. Furthermore, Coxon supported temporary restrictions on capability improvements while governments develop stronger global safety and accountability standards.

Also Read: India Targets 15 Crypto Platforms Over Anti-Money Laundering Compliance Failures