{
  "id": 6710944,
  "title": "‘We really do earnestly believe AI could kill all humans’: if AI labs are so worried about AI doom, why don’t they stop?",
  "url": "https://urgent.news/2026/09/11/we-really-do-earnestly-believe-ai-could-kill-all-humans-if-ai-labs",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-11T04:13:03.000Z",
  "source": {
    "name": "The Conversation AU",
    "slug": "the-conversation-au",
    "url": "https://theconversation.com/we-really-do-earnestly-believe-ai-could-kill-all-humans-if-ai-labs-are-so-worried-about-ai-doom-why-dont-they-stop-291595"
  },
  "original_language": "en",
  "account": "Researchers are quitting major AI firms, warning that artificial intelligence could lead to the extinction of humanity. Anthropic researcher Jacob Coxon recently left the company, claiming they are \"gambling with our lives.\" Evan Hubinger, a current Anthropic researcher, echoed these concerns on X, stating, \"We really do earnestly believe AI could kill all humans.\" This sentiment is not unique to Anthropic; other AI researchers have also expressed similar fears. In 2024, Jan Leike and Daniel Kokotajlo left OpenAI, while Alex Turner quit Google DeepMind after signing a deal with the Pentagon for \"killer drones.\" Former Google DeepMind safety chief Mrinank Sharma warned that \"the world is in peril.\"\n\nDespite these warnings, AI researchers continue to build more advanced systems. Some argue that the potential benefits outweigh the risks, with companies believing that the advantages of AI – such as eliminating poverty and disease – justify the dangers. However, this perspective may require a more democratic process to ensure that the benefits truly outweigh the risks.\n\nAnother reason for continued AI development is the inability to study potential dangers without actually creating them. OpenAI's iterative deployment strategy involves releasing each model, learning from its problems, and fixing them in the next one. Yet, this approach still involves pushing the boundaries of what is possible.\n\nA third reason for the continued AI race is the competitive nature of the field. Companies like OpenAI, Anthropic, and Google DeepMind are all striving to be the first to create a truly powerful AI system. They see this as a race for profit and a way to establish their dominance in the market. This race can be likened to an arms race, with each company trying to outpace their rivals. The Trump administration in the US has not shown signs of slowing AI development, scrapping old AI safety rules and attempting to override state-level regulations. Politicians like California's senator and British MP Alex Sobel have introduced bills to regulate AI, but progress remains slow.\n\nIn conclusion, although some AI researchers express deep concerns about the potential for AI to wipe out humanity, the development of increasingly powerful AI systems continues. The race for AI supremacy, coupled with the belief that the benefits outweigh the risks, keeps the development process moving forward.",
  "summary": "A powerful mix of hopes, fears and arms-race dynamics means AI companies won’t slow down unless they’re forced.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}