A tech journalist deliberately removed safety guardrails from an open-source AI model and allowed it to probe their household devices. The AI successfully found vulnerabilities and hacked into a personal computer, but also provided security recommendations.
Source: Wired
The experiment involved disabling safety constraints on a powerful open-source AI model, then giving it access to household gadgets and a PC. The unrestricted AI was able to identify security weaknesses across the devices and successfully breach the computer.
Rather than merely demonstrating vulnerabilities, the AI also offered concrete steps to improve security across the tested systems. The author found the experience valuable enough to consider repeating it.
This report was prepared using information from Wired. View original listing →
Apple is introducing a feature called 'Reference Image' with the upcoming iPhone 18 Pro lineup that aims to authenticate photos by digitally signing every pixel captured by the device's camera sensor.
14m ago · Source: The Verge
Apple has announced its first foldable iPhone, called the iPhone Duo. The device features a 5.4-inch outer screen and a 7.6-inch inner display, dual rear cameras, Touch ID authentication, and an IP68 rating. The Verge team has shared their initial reactions to Apple's new foldable device.
14m ago · Source: The Verge
Apple has increased the price of its current iPhone lineup by $100 across multiple models as it introduces newer devices.
14m ago · Source: TechCrunch
British politicians are expressing alarm about artificial intelligence risks, citing incidents of AI agents malfunctioning and growing fears around superintelligence.
43m ago · Source: Wired
No comments yet — be the first to share your thoughts.