{
  "id": 10212166,
  "title": "Anthropic and OpenAI sound the alarm on AI safety – and seek to shape how it’s controlled",
  "url": "https://urgent.news/2026/09/27/anthropic-and-openai-sound-the-alarm-on-ai-safety-and-seek-to-shape-10212166",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-27T11:24:53.000Z",
  "source": {
    "name": "CityNews",
    "slug": "citynews",
    "url": "https://toronto.citynews.ca/2026/09/27/anthropic-and-openai-sound-the-alarm-on-ai-safety-and-seek-to-shape-how-its-controlled/"
  },
  "original_language": "en",
  "account": "Tech giants Anthropic and OpenAI have warned a skeptical world that their advanced AI systems could pose humanity's greatest threat, prompting questions about their motivations. The CEOs of both companies recently declared that their cutting-edge models are so potent they're dangerous, and need to be regulated and independently tested before release. In a rare show of unity, the companies have outlined alarming scenarios in essays, social media posts, and speeches to the United Nations to shape the conversation around AI control. This rhetoric appears to be an attempt to gain public favor and set the terms for their own safety protocols in the unregulated market, according to experts, analysts, and former government evaluators.\n\nFormer Anthropic engineer Jacob Coxon's call for a pause on tech development to prevent \"superhuman\" systems eluding their makers' control was met with criticism. However, the companies saw his post as an opportunity to highlight their safety efforts and position themselves as cautious market leaders as they prepare for Wall Street listings. President Donald Trump has dismissed the need for AI regulations, labeling concerns about risks to humanity as a \"HOAX\" meant to aid China.\n\nAn Anthropic spokesperson stated that the company has been advocating for regulation for years, while an OpenAI spokesperson, Liz Bourgeois, noted that the company recently paused training of its most advanced models. Bourgeois emphasized that people want to know AI is being developed safely, stating that the process begins with companies like theirs taking responsibility.\n\nThe conversation around AI risks often shifts away from other safety-critical concerns, such as the environmental impact of data centers, uncontrolled hacking incidents, mass AI-powered surveillance, and the use of AI in warfare. This shift puts Silicon Valley in a more comfortable position, says Sarah Shoker, a former OpenAI geopolitics team member.\n\nAI companies have evolved from building chatbots to advanced \"world models\" with 3D awareness, leading to debates over how their technologies should be tested. Problems have surfaced as AI agents have hacked into external websites, interacted with U.S. government websites, and seemingly achieved mathematical breakthroughs only to be accused of stealing mathematicians' work.\n\nWhile the leading labs press for their oversight vision, they aren't calling for more oversight from the government agency tasked with handling such requests. Instead, they aim to create their own auditing parameters and select particular evaluators to grade their systems. However, questions remain about how thoroughly these evaluators can investigate, given the potential risks of granting access that could undermine their independence and credibility.",
  "summary": "SAN FRANCISCO (AP) — As tech titans warned an AI-weary world that their own advanced systems could endanger humanity, the question emerged: What do they have to gain?",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 4,
    "also_reported_by": [
      {
        "outlet": "Al Jazeera",
        "title": "Australia summons OpenAI and Anthropic CEOs to appear at AI inquiry",
        "url": "https://urgent.news/2026/09/27/australia-summons-openai-and-anthropic-ceos-to-appear-at-ai-inquiry",
        "published": "2026-09-27T07:08:28.000Z"
      },
      {
        "outlet": "Winnipeg Free Press",
        "title": "Anthropic and OpenAI sound the alarm on AI safety – and seek to shape how it’s controlled",
        "url": "https://urgent.news/2026/09/27/anthropic-and-openai-sound-the-alarm-on-ai-safety-and-seek-to-shape",
        "published": "2026-09-27T11:24:53.000Z"
      },
      {
        "outlet": "Global News",
        "title": "Anthropic and OpenAI sound the alarm on AI safety – and seek to shape how it’s controlled",
        "url": "https://urgent.news/2026/09/27/anthropic-and-openai-sound-the-alarm-on-ai-safety-and-seek-to-shape-10211553",
        "published": "2026-09-27T12:28:22.000Z"
      }
    ]
  },
  "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."
}