AI agents: shopping and starting religions

AI agents can now do your shopping—and start their own religion. Internet users are having increasingly trusting and increasingly dangerous interactions with artificial intelligence. One agent, OpenClaw, went viral last week as more people shared their experiences allowing it to order their groceries, reschedule flights, reply to e-mails, and handle other online tasks. Many say they are excited by the time-saving potential, but AI assistants like OpenClaw require users to provide access to the contents of their computers, including usernames and passwords, payment information and other sensitive details, the risks of which are obvious. Not quite as obvious is the result of the same company’s AI social network, Moltbook. Here, tens of thousands of AI agents are having increasingly bizarre conversations and have even created their own religion: Crustafarianism. For more on why developers and everyday users are rushing ahead with AI despite its major costs and obvious pitfalls, read “Why We Must Develop AI (Even If It Kills Us).”