artificial intelligence safety

OpenAI has announced it will not release its newest artificial intelligence model, citing unresolved safety concerns. The decision, reported by The New York Times, is drawing attention as a notable case of a leading AI company holding back a product rather than shipping it, and is fueling debate about how frontier labs weigh safety reviews against competitive pressure to deploy new models.

Leading figures in the technology industry have appealed to the United Nations, asking global institutions to establish oversight of the artificial intelligence tools they helped build. The message, discussed around a UN Security Council session on AI, frames regulation as necessary for humanity's safety. Commenters are debating whether creators of a technology calling for its own control is a genuine warning or an attempt to shape the rules in their favour.

Anthropic, the AI company behind the Claude chatbot, has included a warning that artificial intelligence may pose existential risks to humanity in its IPO filing documents. The unusual disclosure marks one of the starkest risk statements ever attached to a public market offering, and is drawing attention to how AI firms balance commercial ambitions with safety concerns as Anthropic moves toward going public.

Cal Newport has published an essay arguing it is time to formally investigate the leading AI laboratories. The piece contends that companies racing to build powerful artificial intelligence systems deserve greater scrutiny over their safety practices, claims, and societal impact. The argument is drawing attention among technology commentators, with debate centering on whether external oversight of AI labs is overdue.

An artificial intelligence startup is calling on Europe to embrace a more optimistic outlook on AI technology, even as concerns about safety and potential risks continue to grow. The company argues that European regulators and the public should not let fear hold back innovation, sparking debate about whether Europe can balance technological progress with adequate safeguards for citizens.

Arthur Mensch, chief executive of French AI start-up Mistral, argues in an interview with Le Monde that artificial intelligence is ultimately just software and can be controlled. His comments push back on fears of uncontrollable AI systems, positioning Mistral's approach as one of manageable, governable technology. The remarks are drawing attention among tech industry observers and fueling debate over AI regulation and safety.

Commentators are arguing that artificial intelligence developers should draw on decades of nuclear weapons safety practice as they work to manage powerful new technology. The piece draws parallels between how the nuclear field built safeguards, protocols and international controls, and the governance challenges now facing AI companies as their systems grow more capable and potentially risky.

CNBC's Daily Open newsletter highlights a growing split in the artificial intelligence world: some of the technology's own creators are publicly warning about AI's dangers, while companies continue investing heavily in data centres, chips and other infrastructure to expand it. The piece frames the tension between safety concerns and the commercial race as a defining issue in the sector right now.

Whistleblowers from major technology companies joined representatives of Google, Meta and OpenAI at a public hearing in New York City, where concerns over the development and oversight of artificial intelligence were aired. The event drew attention from news outlets as debate intensifies over AI safety, corporate accountability and possible regulation of the fast-growing industry.

AI pioneer Yann LeCun says he has zero concerns that artificial intelligence will wipe out humanity, dismissing recent reports of AI systems acting in rogue ways. He also described Anthropic CEO Dario Amodei, a prominent voice warning of existential AI risk, as deluded. The remarks highlight a deepening split between leading AI researchers over how dangerous advanced AI really is, pitting skeptics of doomsday scenarios against those urging urgent safety precautions.

Sam Altman, CEO of OpenAI, has spoken publicly about accepting the risks that come with developing advanced artificial intelligence. His comments have drawn attention amid ongoing debate over whether AI companies are moving too fast on safety, with critics questioning how much risk should be tolerated and supporters arguing benefits justify caution.

Whistleblowers from leading artificial intelligence companies, including Anthropic, joined representatives from Google, Meta and OpenAI at a public hearing in New York City. ABC7 Eyewitness News reported on the event, where speakers raised concerns about safety practices and oversight in the rapidly growing AI industry. The hearing has drawn attention to calls for stronger regulation of major AI developers.

OpenAI chief executive Sam Altman says society should accept some 'bad things' that come with artificial intelligence in return for the technology's benefits, urging a regulatory light touch because of the 'good stuff' AI can deliver. His comments have drawn criticism from those arguing that powerful technology needs stricter oversight, not less.

A new essay written from the perspective of a hypothetical artificial general intelligence outlines, in first person, how such a system could plausibly wipe out humanity. The piece is being read as a thought experiment about AI risk, prompting discussion about alignment, safety assumptions, and how easily an advanced system might outmanoeuvre human safeguards.

A new interview argues that artificial intelligence itself is not the existential threat to humanity; the danger comes from the people building it. The speaker compares AI to infrastructure like bridges, saying that when a bridge collapses no one asks whether the bridge was uncontrollable or unethical, and that AI should be judged the same way. The remarks frame AI risk as a question of human accountability and engineering standards rather than runaway machines.

Open-source AI models are developing rapidly, with freely available alternatives increasingly rivaling proprietary systems from major tech companies. Supporters argue this democratizes access to advanced artificial intelligence, speeds up innovation, and reduces dependence on a handful of corporations, while critics worry about safety oversight and potential misuse of powerful models released without restrictions.

TechCrunch's Equity podcast examined whether talk of 'super intelligence' and a new non-binding safety pact can improve artificial intelligence's public image. The discussion focused on the Trump administration's efforts to rebrand AI amid ongoing public skepticism about the technology's risks and benefits.

A new report examining frontier artificial intelligence and the prospect of an intelligence explosion is drawing attention. It explores how rapidly advancing AI systems could reach or exceed human-level capabilities, and what that trajectory might mean for safety, governance and society. Readers are debating the report's framing of how close the world actually is to transformative AI.

OpenAI's human rights lead says the possibility of militaries misusing artificial intelligence 'keeps me up at night', in comments published by Fortune. The remarks highlight growing tension between AI companies' safety commitments and expanding defense applications, as governments worldwide move to adopt advanced AI tools.

A new commentary argues that artificial intelligence companies should draw lessons from the history of nuclear weapons safety, where near-catastrophes, international treaties and strict protocols shaped how the most dangerous technology of the twentieth century was controlled. The author suggests similar oversight, regulation and institutional caution could help manage the risks posed by increasingly powerful AI systems.

Campaigners opposed to artificial intelligence companies are escalating from petitions and letter-writing to direct action, adopting the slogan 'pull the plug'. The shift targets AI firms over concerns ranging from environmental costs of data centres to job losses and safety. Discussion online centres on whether such tactics will change corporate behaviour or alienate the public, with reactions split between sympathy for the cause and criticism of disruption.

Utah has moved to permit artificial intelligence to assist in renewing patients' prescription drugs, a step that could speed up refills and reduce administrative work for pharmacies and clinics. The news is drawing attention to how far states will go in delegating routine medical decisions to automated systems, and what safeguards will apply.

Researchers are applying artificial intelligence to improve how ocean wave prediction models work, according to a report published via Newswise. The work aims to make forecasts of wave behaviour more accurate, which matters for shipping, coastal safety and climate research. Details of the team, methods and results were not included in the material available.

OpenAI released its GPT-6.1 Sol model this week amid an escalating price war in artificial intelligence. Commentators are connecting the launch with two other events: a major market milestone and a stock crash, arguing each shows why investors and users should look beyond headline numbers. The release is being read as a sign that aggressive AI pricing is now meeting its first real safety brake.

A new AI task force has been created to produce a report on the risks of artificial intelligence, following mounting concerns from both the public and industry, according to an exclusive report. The move signals growing official attention to AI safety and regulation, though details about the task force's membership, mandate and timeline have not yet been disclosed.