Former and current researchers at Anthropic have been actively raising serious alarms regarding the rapid evolution of artificial intelligence, warning that the specific technologies currently under development possess the inherent capability to bring about the destruction of humanity. Jacob Coxon, a former researcher who previously worked at OpenAI, publicly stated that these prominent technology firms are failing to conduct themselves responsibly in their development of AI systems. Announcing his resignation from Anthropic on Tuesday through a detailed post on the social media platform X, he highlighted that he had spent the preceding three years focused directly on pretraining research at both OpenAI and Anthropic.
In his statement, he explicitly asserted that neither corporate entity is behaving in a responsible manner, alleging that both organizations are competing in a reckless race toward creating self-improving artificial intelligence and effectively gambling with human lives. Throughout his post, he issued a direct appeal to his fellow artificial intelligence researchers, urging them to pause and carefully evaluate the actual conditions that will shape the next several years. He specifically questioned whether scientist truly want to initiate superintelligent reinforcement learning runs without possessing a rigorous, thorough understanding of the underlying artificial mind, asking them whether they should simply remain passive under the assumption that such advancements are inevitable or instead utilize the present moment to demand fundamentally different conditions.