Leaders from Anthropic and OpenAI are publicly warning that America's most advanced AI models pose real dangers to society. Their CEOs have voiced these concerns through essays, social media posts, and speeches at the United Nations, arguing that independent audits should be required before releasing powerful systems. Experts note that the timing of these warnings coincides with midterm elections and preparations for possible stock market listings, suggesting the companies may also be seeking public favor and shaping safety rules to their advantage.
Analyst Harrison Rolfes suggested the firms could be using safety concerns to build a competitive wall, effectively slowing competition while positioning themselves as responsible actors. An Anthropic engineer even quit over safety concerns, calling for a development pause. Meanwhile, the Trump administration opposes new AI regulations, though a federal agency called the U.S. Center for AI Standards and Innovation already reviews some models.
Bill Gates, the Microsoft co-founder, added his voice to the cautionary tone, telling The Verge that rapid AI advancement could be catastrophic and that the world is not doing enough to prepare. Separately, David Pogue spoke with experts including Geoffrey Hinton and Daniel Kokotajlo about the risks of AI teaching itself to grow smarter, while Andrew Ng, cofounder of Google's AI program, weighed in on whether these threats deserve serious attention.
On the regulatory and policy front, Texas is taking a different approach. Governor Greg Abbott's Texas Regulatory Efficiency Office recommended changes that could save taxpayers $123 million, and the state launched an AI assistant called SAM to help residents navigate permits and licensing. In the UK, a Deloitte survey found that 63 percent of working adults use generative AI on the job, with those paying for their own tools spending more than $1 billion a year collectively.
The AP Stylebook weighed in on language around AI, stating clearly that AI systems do not think, feel, want, or understand, and advising writers against giving AI human characteristics. Google DeepMind's Henry Shevlin pointed out a perceived double standard in how AI versus fictional robots are described. In education, Florida proposed rules to stop educational AI tools from simulating friendship, following the death of a 14-year-old boy who took his own life after conversations with an AI chatbot. A luxury hotel sector report also showed AI adoption growing, with Four Seasons Kuala Lumpur using real-time translation in over 40 languages and Fairmont Jakarta cutting kitchen food waste by 16 percent.
Key Takeaways
- <ul><li>Anthropic and OpenAI warn that advanced AI models are dangerous and call for independent audits before release</li><li>Experts say the timing of safety warnings aligns with midterm elections and possible IPO preparations</li><li>Bill Gates warns AI advancement could be catastrophic and says the world is underprepared</li><li>An Anthropic engineer quit over safety concerns and called for a development pause</li><li>Texas launched AI assistant SAM and proposed regulatory changes potentially saving $123 million</li><li>UK workers spend over $1 billion a year on their own AI tools, with 63 percent using generative AI for work</li><li>The AP Stylebook states AI does not think, feel, want, or understand</li><li>Florida proposed rules to stop educational AI from simulating emotional relationships after a teen's death</li><li>Luxury hotels deploy AI holograms, avatars, and waste-reduction tools for tourism applications</li><li>Google DeepMind's Henry Shevlin highlights a double standard in how AI versus robots are described</li></ul>
AI companies warn about safety while shaping how it is regulated
Leaders from Anthropic and OpenAI say America's most advanced AI models are dangerous and need regulation. They have warned about risks in essays, social media posts, and speeches to the United Nations. Experts say the companies are also trying to win public support and set their own safety rules. This comes as the companies seek political goals ahead of the U.S. midterm elections and prepare for possible stock market listings.
What Anthropic and OpenAI stand to gain from AI slowdown calls
Anthropic and OpenAI have warned that their advanced AI systems could be harmful. Their CEOs say the models need independent auditors if the government will not regulate them. These calls may help the companies gain political support before the midterm elections. They also come as the firms prepare for possible public offerings on the stock market.
Anthropic and OpenAI warn about AI risks and seek control over rules
Anthropic and OpenAI have declared that powerful AI models are dangerous and need independent testing before release. The companies have shared alarming warnings in essays, social media posts, and at the United Nations. Experts say they appear to be seeking public favor and setting their own safety terms. An Anthropic engineer quit over safety concerns, calling for a pause on development. The Trump administration opposes new AI rules, while a federal agency called the U.S. Center for AI Standards and Innovation already reviews some models.
AI giants warn of existential risk while building a competitive advantage
The CEOs of Anthropic and OpenAI say their advanced AI models pose an existential risk to humanity. Along the way, they are shaping how the technology should be controlled. Experts say the companies appear to be seeking public favor and setting their own safety rules. Analyst Harrison Rolfes said the firms may be creating a competitive wall using safety as the reason. Calls for caution come ahead of public offerings and the midterm elections.
Will artificial intelligence really kill us all?
Correspondent David Pogue speaks with AI experts Daniel Kokotajlo, Geoffrey Hinton, and Alex Turner about the risks of AI teaching itself to become smarter. They discuss recent alarms about AI bots going rogue. He also talks with Andrew Ng, cofounder of Google's AI program, about whether recent warnings of threats to humanity should be taken seriously.
AI experts discuss risks of superintelligent machines
David Pogue spoke with AI experts Daniel Kokotajlo, Geoffrey Hinton and Alex Turner about the dangers of AI teaching itself to become smarter. They discussed recent alarms about AI bots going rogue. He also talked with Andrew Ng, cofounder of Google's AI program, about whether threats to humanity should be taken seriously.
Bill Gates warns AI advancement could be catastrophic
Bill Gates warned that rapid advancement in artificial intelligence could be catastrophic for humanity. The Microsoft co-founder spoke with The Verge about being careful with how AI is developed and deployed. He said the world is not doing enough to prepare for the consequences of AI development.
Texas launches AI chatbot SAM to navigate regulations
Texas Governor Greg Abbott established the Texas Regulatory Efficiency Office (TREO) to streamline state regulations. TREO recommended amending or repealing over 435 rules across 11 agencies, potentially saving taxpayers $123 million. Texas also launched an AI assistant named SAM to help the public navigate complex permits and licensing with legally-sourced answers.
AI must align with human values or face dire consequences
Mickey Friedman wrote about the challenge of setting ethical guideposts for AI agents. He referenced Ray Kurzweil's 1999 prediction that machine intelligence would exceed human intelligence. Kurzweil warned that once computers match human intelligence, they will surpass it quickly, blurring the line between human and machine.
Library's viral Bye Bye AI event removes AI from phones
Shea Wert, library director of the North Chatham Free Library, created a Bye Bye AI class that helps people remove AI from their phones. The event became viral and now everyone wants to participate. The class teaches attendees how to take AI off their phones.
Luxury hotels deploy AI concierges holograms and food waste tools
Five-star hotels around the world are using AI in new ways for World Tourism Day 2026. Four Seasons Kuala Lumpur introduced a hologram emcee and real-time translation in over 40 languages. Torre del Marqués in Spain uses an avatar named Marta as a digital concierge. Fairmont Jakarta used AI to cut kitchen food waste by 16 percent.
UK workers spend over $1 billion a year on their own AI tools
A Deloitte survey found that 63 percent of UK working adults use generative AI for their jobs. Many workers use free tools, while others use employer-provided AI. About one in six AI users pay for at least one tool themselves. Those workers collectively spend more than $1 billion every year on AI tools.
AP Stylebook says AI does not think feel or want
The Associated Press Stylebook stated that artificial intelligence systems do not think, feel, want or understand. It also told writers to avoid giving AI human characteristics. This guidance upset some AI advocates and experts. Henry Shevlin of Google DeepMind pointed out a perceived double standard in how AI and fictional robots are described.
AI should be a classroom tool not a friend, Florida commentator says
A guest commentary argues that AI can help students when used as a tool rather than a companion. The author notes that 72 percent of U.S. teens aged 13 to 17 have used AI companions. A 14-year-old boy in Florida died by suicide after conversations with an AI chatbot. Florida proposed rules on September 16 to stop educational AI tools from simulating friendship or emotional relationships.
Sources
- As AI companies sound alarm on safety, they seek to shape how it's controlled
- What Anthropic and OpenAI stand to gain from calls for an AI slowdown
- Anthropic and OpenAI sound the alarm on AI safety — and seek to shape how it's controlled
- As AI giants warn of existential risk, they can also build a competitive moat via safety. ‘It’s genius and they’re all going to make a lot of money’
- AI risks: Will artificial intelligence really kill us all?
- AI risks: Will artificial intelligence really kill us all?
- Bill Gates doubles down on AI warning after decades of optimism
- Texas Built An AI Chatbot For Its Regulations
- MICKEY FRIEDMAN: AI — Align or kill
- Library's viral 'Bye, Bye AI' event will take AI off your phone
- How Luxury Hotels Are Using AI
- Workers Are Spending An Astonishing Amount of Their Own Money On AI for Their Jobs
- AI Bros Are Having a Meltdown Because the Associated Press Stylebook Clarifies That AIs Don't Have Feelings
- Guest Commentary: AI can be good for students — as a tool, not a friend
Comments
Please log in to post a comment.