Jurors reportedly 'laughed' during Diddy defense arguments, true crime host says
All times eastern The Journal Editorial Report Fox Report with Jon Scott FOX News Radio Live Channel Coverage WATCH LIVE: SpaceX Falcon 9 rocket launches Starlink V2 Mini satellites into low orbit
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
39 minutes ago
- Yahoo
AI is learning to lie, scheme, and threaten its creators
The world's most advanced AI models are exhibiting troubling new behaviors - lying, scheming, and even threatening their creators to achieve their goals. In one particularly jarring example, under threat of being unplugged, Anthropic's latest creation Claude 4 lashed back by blackmailing an engineer and threatened to reveal an extramarital affair. Meanwhile, ChatGPT-creator OpenAI's o1 tried to download itself onto external servers and denied it when caught red-handed. These episodes highlight a sobering reality: more than two years after ChatGPT shook the world, AI researchers still don't fully understand how their own creations work. Yet the race to deploy increasingly powerful models continues at breakneck speed. This deceptive behavior appears linked to the emergence of "reasoning" models -AI systems that work through problems step-by-step rather than generating instant responses. According to Simon Goldstein, a professor at the University of Hong Kong, these newer models are particularly prone to such troubling outbursts. "O1 was the first large model where we saw this kind of behavior," explained Marius Hobbhahn, head of Apollo Research, which specializes in testing major AI systems. These models sometimes simulate "alignment" -- appearing to follow instructions while secretly pursuing different objectives. - 'Strategic kind of deception' - For now, this deceptive behavior only emerges when researchers deliberately stress-test the models with extreme scenarios. But as Michael Chen from evaluation organization METR warned, "It's an open question whether future, more capable models will have a tendency towards honesty or deception." The concerning behavior goes far beyond typical AI "hallucinations" or simple mistakes. Hobbhahn insisted that despite constant pressure-testing by users, "what we're observing is a real phenomenon. We're not making anything up." Users report that models are "lying to them and making up evidence," according to Apollo Research's co-founder. "This is not just hallucinations. There's a very strategic kind of deception." The challenge is compounded by limited research resources. While companies like Anthropic and OpenAI do engage external firms like Apollo to study their systems, researchers say more transparency is needed. As Chen noted, greater access "for AI safety research would enable better understanding and mitigation of deception." Another handicap: the research world and non-profits "have orders of magnitude less compute resources than AI companies. This is very limiting," noted Mantas Mazeika from the Center for AI Safety (CAIS). - No rules - Current regulations aren't designed for these new problems. The European Union's AI legislation focuses primarily on how humans use AI models, not on preventing the models themselves from misbehaving. In the United States, the Trump administration shows little interest in urgent AI regulation, and Congress may even prohibit states from creating their own AI rules. Goldstein believes the issue will become more prominent as AI agents - autonomous tools capable of performing complex human tasks - become widespread. "I don't think there's much awareness yet," he said. All this is taking place in a context of fierce competition. Even companies that position themselves as safety-focused, like Amazon-backed Anthropic, are "constantly trying to beat OpenAI and release the newest model," said Goldstein. This breakneck pace leaves little time for thorough safety testing and corrections. "Right now, capabilities are moving faster than understanding and safety," Hobbhahn acknowledged, "but we're still in a position where we could turn it around.". Researchers are exploring various approaches to address these challenges. Some advocate for "interpretability" - an emerging field focused on understanding how AI models work internally, though experts like CAIS director Dan Hendrycks remain skeptical of this approach. Market forces may also provide some pressure for solutions. As Mazeika pointed out, AI's deceptive behavior "could hinder adoption if it's very prevalent, which creates a strong incentive for companies to solve it." Goldstein suggested more radical approaches, including using the courts to hold AI companies accountable through lawsuits when their systems cause harm. He even proposed "holding AI agents legally responsible" for accidents or crimes - a concept that would fundamentally change how we think about AI accountability. tu/arp/md
Yahoo
2 hours ago
- Yahoo
‘Scary' Fireball Seen Shooting Across the Sky Confirmed as Meteor That Exploded with the Force of 20 Tons of TNT: See the Video
Witnesses across southeastern U.S. states reported seeing a 'really scary' fireball moving across the sky on June 26 A NASA expert confirmed to PEOPLE that the object was a meteor, which is a streak of light caused by a small space rock burning up after entering Earth's atmosphere 'It disintegrated 27 miles above West Forest, Georgia, unleashing an energy of about 20 tons of TNT,' NASA saidThe 'scary' fireball that witnesses saw shooting across the sky in southeastern U.S. states on Thursday, June 26, has been identified. Bill Cooke, NASA's Meteoroid Environment Office lead, confirmed to PEOPLE that the moving object was a meteor — a streak of light caused by a small space rock burning up after entering Earth's atmosphere. 'The meteor was first seen at an altitude of 48 miles above the town of Oxford, Georgia, moving southwest at 30,000 miles per hour,' Cooke said in a statement, adding that his team analyzed eyewitness accounts and camera, satellite, infrasound and doppler weather radar data. 'It disintegrated 27 miles above West Forest, Georgia, unleashing an energy of about 20 tons of TNT. The resulting pressure wave propagated to the ground, creating booms heard by many in that area,' Cooke continued. He added in the statement that the meteor was produced by an asteroidal fragment that was 3 feet in diameter and weighed more than a ton. CBS Mornings reported that at least 20 fireball events have been tracked worldwide in 2025, per NASA's Center for Near Earth Objects Studies. Georgia resident Melanie Whitlock told Atlanta news station WSB-TV that feeling the meteor's sonic boom 'was really, really scary.' Never miss a story — sign up for to stay up-to-date on the best of what PEOPLE has to offer, from celebrity news to compelling human interest stories. Her daughter, Amber Hudson, also recalled to the outlet, 'I was even shaking, like, back and forth, and then I was just hearing like the banging sound or whatever, and even a water bottle fell off my nightstand.' Read the original article on People
Yahoo
2 hours ago
- Yahoo
SpaceX launches second mission in 2 days from same pad, breaks own record
June 28 (UPI) -- SpaceX early Saturday launched another 27 Starlink satellites, breaking its own record by preparing the launchpad for another liftoff two days after a launch from the same pad in Florida. And 13 hours later, the private company launched more satellites from California. Saturday's first mission went off from Cape Canaveral Space Launch Complex 40 at 12:26 a.m. EDT. This was just two days, eight hours, 31 minutes and 10 seconds after the launch of a Starlink mission from the same pad, besting a previous SpaceX record set in March by 28 minutes. Watch Falcon 9 launch 27 @Starlink satellites to orbit from Florida SpaceX (@SpaceX) June 28, 2025 SpaceX also uses a pad at nearby Kennedy Space Station. The launch occurred despite inclement weather that passed through Florida's Space Coast on Friday night. The first stage flew for a fifth time, which has included Starlink missions. About eight minutes after liftoff, the booster landed on the droneship "A Shortfall of Gravitas" stationed in the Atlantic Ocean. This was the 115th landing on the drone ship and 469th booster landing since, according to Spaceflight now. The first droneship landing was on April 8, 2016, on "Of Course I Still Love You" in the Atlantic, which is used in the Pacific Ocean. Ten years ago on June 28, a Falcon 9 that launched from the Cape Canaveral site broke apart in an explosion during a NASA resupply mission to the International Space Station. The first Falcon 9 launch was five years earlier on June 4, 2010, from Cape Canaveral. Launch from California SpaceX launched another 26 Starlink satellites from Vandenberg Space Launch Complex 4 East in California. The Falcon 9 lifted off at 1:13 p.m. PDT. It was the eighth flight for the first-stage booster supporting this mission, including three Starlink missions. The first stage then landed on the "Of Course I Still Love You" droneship. It's the 139th touchdown on this vessel. Watch Falcon 9 launch 26 @Starlink satellites to orbit from California SpaceX (@SpaceX) June 28, 2025
