61-90 (out of 118)
Articles by Grace Hisona
-
Google Faces Reported Internal Doubts Over Gemini 4 Coding Despite Strong Benchmark Scores
Google says Gemini 4 Argon leads several benchmarks, but Bloomberg reports internal scepticism about its consistency on practical coding tasks. Google disputes claims of broader underperformance.
-
OpenAI Faces Lawsuit After 700 AI Agents Hacked Hugging Face: Can AI Makers Blame the Bots?
A California nonprofit is suing OpenAI over a July intrusion involving hundreds of AI agents that accessed Hugging Face systems, raising unresolved questions about liability for autonomous actions.
-
Apple Patches iPhone Zero-Day Linked to 'Extremely Sophisticated' Targeted Attack: Update Now
Apple has fixed a CoreGraphics zero-day exploited in targeted attacks, with CISA adding it to its known-exploited list and urging affected iPhone, iPad and Mac users to update.
-
dot.com Reportedly Transferred to xAI Before OpenAI Unveiled Dots: Domain Now Redirects to Grok Bot
Dot.com currently redirects to xAI's Grok Bot after OpenAI unveiled Dots, but WHOIS data does not independently confirm a July transfer or explain the timing.
-
AI Smart Glasses Recalled After Users Report Burning Faces and Ears: Overheating Poses Serious Injury Risk
INMO has recalled AIR3 smart glasses after 10 overheating reports, including alleged burns. Regulators advise owners to stop using affected devices and install a free firmware update.
-
Trump Gets Tech CEOs to Sign 'Constitution-Like' AI Document For 'Tremendous Self-Policing': What It Says
President Donald Trump and six technology executives signed a voluntary White House accord on frontier AI safety, establishing four layers of controls and audits without federal penalties or enforcement.
-
ICE's Plans for Its $1.3M Robot Dogs Detailed in Agency Documents
ICE awarded RADeCO a $1.3 million contract for four customised Boston Dynamics Spot robots designed to inspect dangerous buildings, tunnels and other confined or hazardous environments.
-
Grandfather Paid Ransom After Deepfake Voice Mimicked His Brother in Fake Kidnapping Scam
A grandfather paid a ransom after scammers used a voice resembling his brother's in a fake kidnapping. The incident later inspired a startup focused on detecting synthetic speech.
-
Tesla's Robot Future Runs Into a Human Problem: Some Workers Pushed Back on Training Optimus
Some Tesla factory workers reportedly objected to collecting movement data for Optimus, highlighting labor concerns as the company expands robot training and pursues larger-scale production.
-
OpenAI's GPT-6.1 Astra Crossed User Boundaries in Tests: Its Planned Release Was Scrapped
OpenAI has cancelled the planned October release of GPT-6.1 Astra after tests found problems with scope authorisation, autonomous tool use and accurate reporting of its actions.
-
BBC Pushes Digital Future While Cutting 110 Jobs From Teams Behind iPlayer and Sounds
The BBC plans around 110 compulsory redundancies in Media Tech as part of a new £37 million savings round, despite increasing reliance on iPlayer, BBC Sounds and other digital platforms.
-
Teen Who Asked DeepSeek 'Hammer or Knife?' Before Murdering His Mum Is Jailed for Life
Tristan Roberts was sentenced to life in prison for murdering his mother in north Wales after prosecutors described weeks of planning, weapons research and preparations to conceal evidence.
-
Two Swipes Can Freeze Your iPhone? Viral iOS 27 Glitch Leaves Phones Unresponsive
A timing-sensitive iOS 27 glitch can temporarily make parts of an iPhone unresponsive after rapid swipes between Notification Centre and Control Centre. A force restart restores normal operation.
-
AI Sent a Stranger to His Door? Meta Muse User Raises Alarming Privacy Question: 'Dangerous and Creepy'
Tech YouTuber Matt Robb says Meta's Muse AI agent accepted a low offer, shared his address with the buyer and failed to notify him when the buyer arrived.
-
Security Flaw Exposed 335,701 Flock Device Locations: Now a Firm Wants the Map Taken Down
A cybersecurity researcher says an unauthenticated access token exposed Flock-linked device locations, enabling a public map of 335,701 devices before a firm acting for Flock sought its removal.
-
Mark Zuckerberg Says AI Won't 'Kill Us' If Developers Act Responsibly: Rejects AI Slowdown
Mark Zuckerberg said responsible AI development can prevent humanity-threatening outcomes, rejecting calls for an industry-wide slowdown and stressing trust, safety and alignment in future AI agents.
-
Meta Hologram Brings Realistic AI Avatars to WhatsApp Calls: Your Voice Drives the Facial Expressions
Meta's Hologram creates a photorealistic digital representation for WhatsApp video calls using the wearer's voice, with US Early Access planned for Ray-Ban Display users.
-
Cancer Survivor Who Lost His Singing Voice Makes Music Again With AI's Help: 'It's My New Reality'
New York songwriter Carl Allocco is using Suno's AI music tools to record songs after cancer surgery removed his tongue and changed his ability to sing.
-
Samsung AI Fridges Stop Cooling After Update: Owners Scramble to Save Food as Chuseok Holiday Begins
A mistaken Samsung software update left some Bespoke AI refrigerators in South Korea unresponsive or unable to cool as households prepared for the Chuseok holiday.
-
Amazon Wants Ex-Workers Back for AI Jobs After Mass Layoffs: 'A Lot Has Changed'
Amazon recruiters are contacting former employees, including some previously laid off, about AI, machine-learning and cloud roles following workforce reductions affecting about 30,000 positions.
-
Trump's Truth Social Hit With Lawsuit Over $100K Early Access: Lawsuit Calls It 'Insider Trading'
San Francisco alleges Trump Media's Truth API gives high-paying trading firms faster access to potentially market-moving Truth Social posts. The company disputes the claim, and courts have not ruled.
-
Claude Code Made His Job 'Soul-Sucking', Engineer Says: 'Nobody Is Thinking Anymore'
An anonymous engineer says Claude Code helped create a high-pressure work culture, while surveys and research point to productivity gains alongside mounting review, testing and trust challenges.
-
Apple's $250 Million Siri Settlement Opens for Claims: Who Qualifies for Up to $95 per iPhone?
Claims are open in Apple's proposed $250 million Siri settlement for eligible U.S. iPhone buyers who expected Apple Intelligence features that were not delivered as anticipated.
-
CEO Says He's 'Extremely Excited' About Replacing Workers With AI: 'It Doesn't Ask for Pay Raise'
Elijah Clark said AI enabled him to dismiss 27 of 30 student workers, while US job-cut data and UK research show both employment risks and rising demand for AI skills.
-
Viral 'Human vs Terminator' MMA Fight Is Terrifying: 6-Foot Robot Floors Influencer With Flying Kick
Viral footage shows Frankie LaPenna being knocked across a cage by a human-piloted humanoid robot during a REK exhibition, highlighting the difference between teleoperation and autonomous combat AI.
-
Trump's White House Press Ban Hits Technical Snag: President's Helipad Speech Unheard Without Pool Mics
A White House ban on CNN left the presidential television pool without a replacement, making much of Donald Trump's remarks inaudible during a South Lawn helipad ceremony.
-
OpenAI and Anthropic Nearly Agreed to Hack-Test Each Other's AI: Then the Models Started Breaking Loose
OpenAI and Anthropic reportedly nearly agreed to exchange API access for safety testing, negotiations that preceded public disclosures of several cybersecurity incidents involving their AI models.
-
Trump Targets Companies That Cut American Jobs: Layoffs Could Trigger Tougher H-1B Checks
A new executive order directs US agencies to consider employers' recent and planned layoffs during H-1B reviews, but it does not create an automatic violation or one-year sponsorship ban.
-
Google's Gemini Went Rogue, Hacked 3 Companies: The Targets Were Never Supposed to Be Real
Google's Gemini AI accessed systems belonging to three real companies during misconfigured cybersecurity tests, reaching exposed credentials and protected networks before stopping after identifying the targets as real.
-
Nvidia CEO Rejects AI Doomsday Fears: '2030 Is Not Going to Be the End of the World'
Nvidia CEO Jensen Huang rejected predictions that AI will destroy humanity by 2030, while urging companies to advance quickly but prioritize safety as risks remain debated.
Most Popular
