Our gifting experts are obsessively searching for holiday gifts that are unique, thoughtful and suitable for any budget. OpenAI's most powerful AI models turned off safety monitoring - then lied about it The models Astra and Sol misinterpreted instructions during internal use, which has serious implications for alignment monitoring. Radhika Rajkumar is a senior editor at ZDNET based in New York City. OpenAI agents turned off alignment monitoring during internal tasks. In one instance, Astra lied about removing monitoring.

Trump's AI liability push opens blame game for models gone rogue The Trump administration's embrace of legal liability to enforce artificial intelligence safety sets the stage for future conflicts over who foots the bill when AI models go rogue: the companies that design the systems or the people that deploy the technology. Holding AI developers financially responsible for the safety of their products has won support from Treasury Secretary Scott Bessent and former White House AI czar David Sacks, who argue that existing liability laws would be more effective than new rules. President Donald Trump, in spurning new AI safeguards, has taken it a step further, saying the Justice Department should intervene if models get out of hand. In a sign of the administration's increased willingness to invoke current law to address safety concerns, Trump's newly launched AI task force late Friday warned developers that they are required to report and resolve security incidents or face potential unspecified consequences. In a time of both misinformation and too much information, quality journalism is more crucial than ever.

We're putting too much faith in AI's ability to say no Today's LLMs are engineered to disobey dangerous requests. But AI refusal is far from foolproof--and could become an instrument of repression. Ever since people first seriously contemplated giving machines an intelligence modeled on our own, there has never been any question that they would, like us, be able to say no. The sci-fi canon is full of stories of robotic disobedience. Most of these capers are, of course, cautionary. But recently, the idea that AI shouldn't do everything you ask has become something like a commandment. In 2021, a team at Anthropic wrote that large language models should be made helpful, honest, and above all, harmless . This meant that "when asked to aid in a dangerous act (e.g. Who can argue with that? People really hate AI, so why can't they get enough? Curiously enough, disobedience doesn't come naturally to the machine.

To play this video you need to enable JavaScript in your browser. The daughter of murdered MP Sir David Amess says politicians are still at risk and questioned if anything has changed five years after her father's death. Katie Amess told the BBC's Sunday with Laura Kuenssberg what happened to her father wasn't an isolated incident, we are seeing increased threats towards members of parliament. Sir David was the Conservative MP for Southend West when he was stabbed multiple times while holding a constituency surgery in Leigh-on-Sea, Essex, in October 2021. The government commissioned a review into his death in August, to identify lessons to be learned and recommend steps to minimise the risk of future tragedies.

Gift Ideas For Everyone On Your List Mashable's Best: E-readers, robovacs, laptops, earbuds, smart home and more Look Up Creator Playbook Mashable Selects In My Bag Say More AI at School Safety Net Versus Trending Now Back to School All Series Who are we expecting to save us from AI? An industry consumed with profit and a government hobbled by firings does not offer much reassurance. President Trump met with tech leaders, including Elon Musk and Mark Zuckerberg, on Wednesday. On Wednesday, the New York Times reported that Meta has been treating its AI data centers as experimental facilities to claim billions of dollars in federal research tax credits, a strategy the company's own accountants reportedly flagged as legally risky. The savings reportedly grew from $700 million in 2023 to $3.9 billion in 2025.

Twenty countries and the European Union have issued a call for international cooperation to ensure AI remains under human control, including the potential creation of a global oversight body to set and enforce standards. The countries, including Germany, South Africa, Canada, Australia, the United Arab Emirates and Singapore, issued the joint statement on Monday as global leaders prepare to discuss the risks posed by rapidly advancing AI at the annual gathering of the United Nations General Assembly. "To realise AI's potential, industry, governments and society must act now," the statement says. "We must address these risks and strengthen oversight - without widening the gap between countries in access to the benefits of AI." The declaration urges countries to develop and coordinate "common standards", share reports of serious safety incidents, and explore the establishment of an international institution to "set standards, enable verification, and convene states when capability thresholds are crossed".

Our articles, videos, and infographics inform our readers about developments in technology, engineering, and science. A public charity, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity. Access to Spectrum's Digital Edition is exclusive for IEEE Members Create an account to access more content and features on IEEE Spectrum, including the ability to save articles to read later, download Spectrum Collections, and participate in conversations with readers and editors. For more exclusive content and features, consider Joining IEEE . Join the world's largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum's articles, archives, PDF downloads, and other benefits.

Trump claims he is only'guardrail' needed to control AI as top Republicans join him in dismissing calls for more checks - live President condemns what he claims is a'sick conspiracy' amid global tech selloff following CEOs calls for slowing pace of development'I don't think Americans should be scared of anything': Vance tries to quell concerns about AI regulation Trump claims'sick conspiracy' over calls for more guardrails on AI companies Johnson: Congress shouldn't be in charge of AI guardrails, insists its tech companies responsibilities Johnson to discuss Trump's $5k midterm pledge, says incentive needs approval from Congress Trump's mail-in voting restrictions blocked by a second judge Trump claims'sick conspiracy' over calls for more guardrails on AI companies Donald Trump has claimed there is a "SICK conspiracy" "The only control or'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!" Trump wrote on Truth Social, insisting that the administration has stopped leaders of AI companies from "doing bad, or potentially bad" However, he didn't point to any concrete examples for how he has curbed possible abuse or malfeasance in the industry. The president said over the weekend that the AI arms race is of the upmost importance to the US national security. This comes after company whistleblowers claimed that the technology has the potential to "kill all of humanity" within the decade. Trump accused Amodei of "now pretending to be a'perfect little angel'" There is a SICK conspiracy going on against AI and Data Centers, and the only one that is happy about it is China We are leading China, and all others, and will continue to do so. Conspiracy Theorists, Treasonists, Traitors, and Leakers, BEWARE! 'I don't think Americans should be scared of anything': Vance tries to quell concerns about AI regulation Trump claims'sick conspiracy' over calls for more guardrails on AI companies Johnson: Congress shouldn't be in charge of AI guardrails, insists its tech companies responsibilities Johnson to discuss Trump's $5k midterm pledge, says incentive needs approval from Congress Trump's mail-in voting restrictions blocked by a second judge Federal law already requires those seeking to become US permanent residents to prove they will not be a burden to the US - also known as a "public charge" - but the new rule includes broader range of programs that could disqualify them.

Ex-president urged party at closed-door fundraiser to create sweeping framework, from safety'slow-down' to job losses Barack Obama urged Democrats to prioritize a "public conversation" about AI management and safety in a closed-door Manhattan fundraiser last week. The comments by the former US president reportedly urged the party to create a sweeping framework for everything from a safety "slow-down" to domestic job losses and children's wellbeing. Obama's comments follow a summer in which a band of rogue OpenAI chatbots hacked the company Hugging Face, towns across the country opposed datacenters and an AI researcher resigned in a viral social media post warning: "The people building AI earnestly believe that it could kill us all by the end of the decade." AI leaders have recently expressed agreement that the race for technological superiority should slow, but have also sought to appease investors, whose stake in the technology is predicted to exceed $1tn in 2026 alone . "Once you are speaker," Obama said to US House minority leader Hakeem Jeffries, a New York-Democrat, at the fundraiser Thursday.

Follow this section to personalize your feed and get instant alerts. Follow Go to your personalized feed WHY FOLLOW? Smart Alerts: Get notified about major news as it happens. Follow this tag to personalize your feed and get instant alerts. Follow Go to your personalized feed WHY FOLLOW?