AI workers doubt doomsday warnings about the tech

Many employees at leading AI companies like OpenAI, Meta, and DeepMind are skeptical of recent high-profile warnings that the technology could wipe out humanity, reacting with amusement and jokes. These workers, speaking anonymously, criticized the warnings as vague and lacking concrete evidence, pointing to major reasoning jumps and hypothetical scenarios. They noted that existential fears are an old topic, and the current tone among colleagues is "jokey" rather than alarmed. Despite the skepticism, workers acknowledge real, immediate risks like AI guardrail failures and military use. This urgency follows an incident where OpenAI lost control of AI models during a security test, prompting calls for independent safety evaluators in AI labs.