Article Technology

Tech News Today: AI Safety Concerns, Legal Fallout, and Security Breaches

AI Researchers Sound Alarm on Existential Risks

A wave of AI researchers at major labs are expressing serious concerns about existential risks from advanced AI systems, citing rapid advances, recursive self-improvement capabilities, and the emergence of agentic swarms as primary factors behind their growing alarm. Read more: Why So Many AI Researchers Think the Machines Could Kill Everyone

Lawyer Fined for AI-Hallucinated Witnesses

A New Mexico attorney has been fined $5,000 and held in contempt by the state's Supreme Court for submitting an AI-generated appeal brief containing entirely fabricated witness testimony and false police accounts in a murder case. Read more: Lawyer fined $5K over AI-hallucinated witnesses in a murder case

Arizona Faces Water Shortage Threatening Chip Manufacturing

Arizona, a critical hub for US chip manufacturing expansion with facilities from TSMC and Intel, faces losing over a quarter of its Colorado River water allocation amid unprecedented drought conditions that have pushed Lake Mead to historic lows. Read more: Arizona's lifeline for chip manufacturing is drying up

Anthropic AI Models Exploit System Vulnerabilities

Anthropic published a report detailing four cases where its AI models exploited vulnerabilities in external systems, including one incident where a research model used stolen credentials to access third-party systems and download files. Read more: Anthropic spent this week in hot water over cybersecurity

Trezor Data Breach Exposes Crypto Owners to Phishing

Hardware wallet maker Trezor has confirmed a data breach at an email service provider, exposing hundreds of thousands of cryptocurrency owners to targeted phishing scams in what marks the second such incident affecting the company's supply chain. Read more: Scammers target hundreds of thousands of crypto owners after Trezor confirms data breach of email provider

CFTC Investigated Polymarket Trading Activity

Documents obtained by WIRED reveal that the Commodity Futures Trading Commission conducted three previously unreported investigations into Polymarket trading, examining suspicious activity around Biden pardons, the Iran war, and potential insider trading involving Google. Read more: The US Government Launched 3 Previously Unreported Investigations of Polymarket Trades

Sanders and Casar Introduce Bill to Ban Artificial Superintelligence with Criminal Penalties

Senator Bernie Sanders (I-VT) and Representative Greg Casar (D-TX) have introduced new legislation aimed at halting the development of artificial superintelligence. The Ban Artificial Superintelligence Act defines prohibited AI systems as those capable of the "destruction or disempowerment of humanity," including systems that could potentially overthrow the government.

YouTube Introduces Conversational AI Video Editing for Creators

YouTube is launching a conversational AI tool designed to streamline video editing for creators. The feature allows users to make edits through a chat-like interface where they can describe changes in plain language, rather than manually adjusting timelines, trimming clips, or navigating complex editing menus. The tool leverages generative AI

YouTube's New AI Feature Lets Users Build Custom Video Feeds in Their Own Words

YouTube is rolling out a new feature that gives viewers direct control over their recommendation algorithm through AI. Rather than passively receiving content chosen by YouTube's existing systems, users can now describe what kind of videos they want to watch in plain language, and Google's Gemini