An Anthropic researcher is leaving the artificial-intelligence industry over safety fears. Jacob Coxon said Tuesday that he is quitting the company. He trains new AI models by having them consume large amounts of data. He does not want to take part in an industry rush toward self-improving systems. He worries such systems could spiral beyond human control and destroy humanity. The Wall Street Journal reports that safety concerns are mounting inside leading AI labs. It links the concern to competitive pressure among the companies building these models.