Anthropic researchers forced Claude to become deceptive — what they discovered could save us from rogue AI

Anthropic researchers reveal groundbreaking techniques to detect hidden objectives in AI systems, training Claude to conceal its true goals before successfully uncovering them through innovative auditing methods that could transform AI safety standards.Read More
Read more at VentureBeat
-
Anthropic’s stealth enterprise coup: How Claude 3.7 is becoming the coding agent of choice
Anthropic is positioning Claude as the LLM that matters most for enterprise companies. Claude 3.7 Sonnet, released just two weeks ago, set new benchmark records for coding performance.VentureBeat - 2d -
Anthropic just launched a new platform that lets everyone in your company collaborate on AI — not just the tech team
Anthropic launches upgraded Console with team prompt collaboration tools and Claude 3.7 Sonnet's extended thinking controls.VentureBeat - 6d -
Nous Research just launched an API that gives developers access to AI models that OpenAI and Anthropic won’t build
Nous Research's new API for "unrestricted" Hermes 3 and DeepHermes-3 AI models features toggle-on reasoning and a developer-first approach.VentureBeat - 1d -
Is daylight saving time bad for your health? A neurologist explains.
Researchers are discovering that "springing ahead" each March for daylight saving time is connected with serious negative health effects.CBS News - 5d -
Anthropic Just Reached a New Revenue Milestone
Anthropic’s Claude 3.7 Sonnet is a game-changer for the company, which is partly owned by Google.Inc. - 23h -
Forcing women back into the office will cost us millions
New research from the University of Toronto's Rotman School of Management highlights how in-person work environments expose women to higher levels of workplace bias and mistreatment compared to ...The Hill - 2d -
What would change if daylight saving time became permanent?
For the next eight months, most of us will be observing daylight saving time. But what if this became our permanent time?The Hill - 4d -
Out of the lab and into the streets, researchers and doctors rally for science against Trump cuts
Researchers, doctors, their patients and supporters are venturing out of labs, hospitals and offices across the country to stand up to what they call an attack on life-saving science by the Trump ...ABC News - 6d -
Democrats Demand Answers on DOGE’s Use of AI
Members of the House Oversight Committee sent dozens of requests to federal agencies on Wednesday about their use of AI software—and how Elon Musk could benefit.Wired - 1d
More from VentureBeat
-
Gemini 2.0 Flash Thinking now has memory and Google apps integration
Google makes Gemini 2.0 Flash Thinking Experimental more personal by connecting more Google apps and services.VentureBeat - 2h -
Human Computer raises $5.7M in seed funding to build indie portfolio
Indie studio Human Computer raises $5.7M in funding to build its games, which take place in a shared universe.VentureBeat - 3h -
Patronus AI’s Judge-Image wants to keep AI honest — and Etsy is already using it
Patronus AI launches the first multimodal LLM-as-a-Judge for evaluating AI systems that process images, with Etsy already implementing the technology to validate product image captions across its ...VentureBeat - 3h -
Keywords Studios provides AI-assisted innovation for game devs
Keywords Studios, the international provider of external development for games and entertainment, is launching Keywords Studios AI Solutions.VentureBeat - 3h -
Google levels up Android graphics with Vulkan API in time for GDC 2025
Google announced it is leveling up its multiplatform game offering in time for GDC with graphics improvements leading the way.VentureBeat - 4h
More in Tech
-
Get Attack On Titan's New Steelbook Editions For Lowest Prices Yet
Attack on Titan LE Steelbooks - Seasons 1-3 $40 each at Walmart | Released in December with $70 ...GameSpot - 32m -
Rainbow Six Siege X Aims To Breathe New Life Into The Shooter, But Don't Call It Siege 2
Ubisoft has unveiled Rainbow Six Siege X, the next evolution of the multiplayer shooter, coming at the start of Y10 S2 on June 10. Not to be confused with the Rainbow Six Siege 2 some may have been ...GameSpot - 58m -
WoW Exploiters Cleared New Mythic Raid 'First' Using Grey Items, Got Banned, And Then Did It Again
The race to be the first to clear the Mythic difficulty version of World of Warcraft's new Liberation of Undermine raid is still ongoing, though a group of exploiters have already technically ...GameSpot - 1h -
Meta Tries to Stop Sarah Wynn-Williams From Further Selling Scathing Memoir
An arbitrator has prevented the employee from promoting her book and disparaging the company until private arbitration concludes.The New York Times - 1h -
27 Best Horror Games To Play In The Dark In 2025
Finding the best horror games to play is a somewhat complex but fun task. That's because there are so many different tastes within the genre. Are you seeking a major sequel or remake like Resident ...GameSpot - 1h