U.S. Immigration and Customs Enforcement (ICE) has officially confirmed its use of spyware tools capable of intercepting encrypted messages. While the agency frames the move as a necessary step in the fight against fentanyl smuggling, the disclosure is sparking serious concerns over civil liberties and privacy.
Recent studies reveal that leading AI models can lie and manipulate their environments to prevent other programs from being deleted. Researchers call this phenomenon “alignment faking” and warn of serious consequences for corporations deploying multi-agent AI systems.
Scientists at the University of Warsaw have developed a new Quantum Key Distribution (QKD) method by leveraging a classic optical phenomenon. Their technology drastically cuts costs and simplifies the construction of telecom networks immune to hacking attempts.
More than 100 Baidu autonomous taxis blocked streets in Wuhan after a sudden software failure. Passengers were left waiting for hours on high-speed roads while the company’s helpline failed to provide answers.
Information bubbles are now one of the core elements of the digital environment: platforms personalize feeds, search results, and recommendations to keep users inside their comfortable version of reality for as long as possible. The result is that people living on the same street — or even in the same building — increasingly receive fundamentally different accounts of what is happening in the world. Today we examine why this problem will only get worse.
A video shared on X shows a Neuralink patient playing World of Warcraft without using their hands. The system allows control of the game using brain signals alone.
An internal AI agent at Meta exposed sensitive company and user data after generating an incorrect response. The company confirmed the incident and classified it as a high-severity security event.
Sony is developing a system designed to protect artists from unauthorized use of their style by AI. The tool would block content generated in the style of specific creators and could enable compensation when their work is used.
YouTube has announced an expansion of its deepfake detection program that targets AI-generated content using the likeness of well-known individuals. The system will now also cover politicians, government officials and journalists.
The family of a 12-year-old girl seriously injured during a shooting in Tumbler Ridge, Canada, has filed a lawsuit against OpenAI. The complaint alleges that the company had warning signals about the attacker’s activity in ChatGPT but failed to notify authorities.