Former Ripple Executive Highlights Anthropic Researcher’s Warning Over AI Safety Risks
Yoshikawa highlighted Coxon’s resignation statement as his concerns circulated among developers and technology investors. Coxon, who spent three years conducting neural network pretraining research at OpenAI and Anthropic, argued that both companies were engaged in a race toward increasingly capable artificial intelligence while failing to adequately address the risks associated with that development.
Coxon said he resigned from Anthropic because of his concerns over the direction of advanced AI research. He characterized competition among major AI laboratories as a potentially dangerous race in which pressure to achieve technological leadership could take precedence over caution, transparency and public safety.
The allegations have particular relevance for technology and fintech investors as artificial intelligence becomes increasingly important across the broader digital economy.
AI safety concerns raised by former Anthropic researcher
Coxon’s central argument is that researchers working on advanced AI systems are aware, to varying degrees, of the potential risks posed by increasingly capable models.
He alleged that employees at OpenAI and Anthropic have different levels of awareness regarding the potential consequences of developing highly capable artificial intelligence. In his assessment, many OpenAI employees do not fully understand the possible dangers, while Anthropic employees are more conscious of those risks but remain concerned about falling behind competing AI laboratories.
Coxon argued that this dynamic could create an incentive for researchers to continue accelerating development even when they have reservations about the potential consequences.
His concern is that a slowdown at one company could allow a rival laboratory to gain a technological advantage, creating competitive pressure that makes it more difficult for individual organizations to prioritize safety.
The claims also raise broader questions about how much influence internal AI safety teams have over decisions involving the development of increasingly capable systems.
Calls for stronger international AI safeguards
Coxon has called on engineers and researchers to consider the broader consequences of their work rather than focusing exclusively on technological progress.
He also advocated for international agreements governing the development of advanced artificial intelligence. In addition, Coxon supports temporary limits on improvements in AI capabilities while governments work toward stronger global standards covering safety and accountability.
His comments reflect his own assessment of practices and decision-making at Anthropic and OpenAI. They should not be interpreted as independently established findings.
The account highlighted by Yoshikawa did not include detailed responses from either Anthropic or OpenAI to Coxon’s allegations.
Yoshikawa’s decision to highlight the resignation statement has nevertheless brought the debate over AI safety beyond specialist research communities, drawing attention from audiences that include investors and participants in the technology and financial sectors.
Writer: Marcus RenfieldCrypto Market Analyst & Onchain WriterMarcus Renfield covers cryptocurrency markets with a focus on onchain data, Bitcoin price action, and emerging market narratives. His writing examines how capital flows, network activity, and broader market structure influence short- and medium-term trends.He aims to provide clear, data-informed analysis for readers seeking a deeper understanding of crypto market dynamics.