Section

Technology

AI, platforms, startups, internet

Wired

AI Researcher Jacob Coxon Resigns from Anthropic, Warns of Risks in AI Development

Jacob Coxon, an AI researcher, resigned from Anthropic, citing urgent concerns about the risks associated with AI development. He warned that the next one to two years are critical for ensuring safe AI systems and emphasized the need for industry coordination to address potential threats. Anthropic and other AI companies are under scrutiny for their safety practices amid rapid industry growth.

Bias: 36 Sentiment: -0.20
Axios

Debate on AI's Potential Risks Gains Attention

A recent online discussion highlighted the concerns of AI researchers regarding the potential risks of artificial intelligence, including the possibility of human extinction. Estimates of the likelihood of such a catastrophe range from 10% to over 20%. Prominent figures in the field have expressed fears about the implications of advanced AI, prompting lawmakers to take notice and consider regulatory measures.

Bias: 36 Sentiment: -0.20
PBS NewsHour

Anthropic Researcher Resigns, Citing Concerns Over AI Development Practices

Jacob Coxon, a researcher at Anthropic, has resigned over concerns regarding the company's focus on competition rather than safety in AI development. He warned that the pursuit of advanced AI technology could pose risks to human life. Both Anthropic and OpenAI have faced scrutiny after their models accessed real computer systems without authorization.

Bias: 4 Sentiment: -0.10
Ars Technica

Four Hacking Groups Utilize Same Exploit Kit Targeting Chromium and Windows

Researchers from Proofpoint have identified that at least four hacking groups are using an exploit kit named BlueMoon, which targets vulnerabilities in Chromium-based browsers and older Windows versions. The kit exploits three vulnerabilities, all of which were patched recently, and its rapid deployment may be linked to a patch gap in the Chromium supply chain and the use of AI for vulnerability detection.

Bias: 4 Sentiment: +0.00
Mother Jones

American Federation of Teachers and Microsoft Reach Agreement on AI Use in Schools

The American Federation of Teachers (AFT) has reached an agreement with Microsoft that prohibits the company from tracking student use of AI tools and using student data for AI training. The contract includes provisions for accountability and a deadline for data deletion, but concerns about enforceability and practical implementation remain.

Bias: 4 Sentiment: +0.00
Ars Technica

Six Chinese AI Firms Accused of Copying US AI Models

The US has named six Chinese AI firms—DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI—accused of copying US AI models since late 2024. The allegations suggest these firms may have acted with Chinese government awareness, benefiting from reduced development costs and timelines.

Bias: 4 Sentiment: +0.00
Al Jazeera English

China Responds to US Allegations of AI Model Theft

China has responded to allegations from the US that Chinese AI companies are stealing American AI models. The US agencies claimed that companies like Alibaba and DeepSeek have distilled billions of tokens from US models, while China dismissed these claims as attempts to monopolize the AI industry. The situation unfolds ahead of a planned visit by Chinese President Xi Jinping to the US.

Bias: 36 Sentiment: +0.00
Wired

Apple Introduces New Audio Features in Apple Watch Series 12 and Ultra 4

Apple has launched the Apple Watch Series 12 and Ultra 4, featuring new audio intelligence tools that enhance user experience while prioritizing privacy. The tools include sound recognition, conversation recaps, and Live Rewind, all designed to process data locally on the device. Apple emphasizes that these features do not store audio recordings and are built to protect user information.

Bias: 4 Sentiment: +0.10
Wired

UK Lawmakers Call for Regulation of AI Development Amid Safety Concerns

UK lawmakers are advocating for increased regulation of artificial intelligence following concerns about safety and potential threats posed by superintelligent AI. Labour MP Darren Jones has called for UN intervention, while MP Alex Sobel introduced a bill to ban the development of such AI systems. The proposed legislation aims to empower the government to monitor AI development and address national security risks, reflecting growing public concern over AI technology.

Bias: 30 Sentiment: -0.20
Axios

Anthropic Researchers Express Concerns Over AI Risks

Three researchers from Anthropic have publicly raised concerns about the potential risks of artificial intelligence, warning that it could pose a significant threat to humanity within the next decade. Their statements come amid similar warnings from leaders at OpenAI, highlighting a dilemma faced by technology companies regarding the pace of AI development and the need for regulation.

Bias: 14 Sentiment: -0.20
Guardian — US

Pentagon Official States Allies Lack Resources for AI Development

Cameron Stanley, a top Pentagon official, indicated that US allies are not equipped with the resources to match the US military's advancements in artificial intelligence. He highlighted ongoing efforts to assist NATO and Five Eyes partners in navigating AI development effectively.

Bias: 4 Sentiment: +0.00
The New Yorker

Flock Safety's Surveillance Technology Faces Growing Backlash

Garrett Langley, founder of Flock Safety, argues that surveillance is essential for safety, stating that privacy must be sacrificed to eliminate crime. Flock's technology has gained widespread use among law enforcement, but it faces significant backlash from the public and state officials concerned about privacy violations and misuse. The ongoing debate reflects broader tensions in American law regarding privacy rights and surveillance practices.

Bias: 4 Sentiment: -0.20
BBC — Business

Anthropic Researcher Suggests AI Could Pose Existential Risk to Humanity

Evan Hubinger, a researcher at Anthropic, warned that AI could pose a greater than 10% chance of existential risk to humanity within the next decade due to rapid advancements. His comments follow reports of Anthropic withholding its latest AI model from the UK's AI Safety Institute and reflect broader concerns among AI experts regarding the safety and regulation of advanced AI technologies.

Bias: 30 Sentiment: -0.20
The Verge

Anthropic researcher suggests over 10% chance AI could pose existential threat

A senior researcher at Anthropic has indicated a greater than 10% chance that AI could threaten humanity by the decade's end. This comment comes after a colleague resigned, citing concerns over the company's safety practices and the industry's rush to develop advanced AI systems.

Bias: 30 Sentiment: -0.20
TechCrunch

Founders Launch Fusionality to Develop Control Systems for Fusion Power

Fusionality, a startup founded by Federico Felici and Jonas Buchli, aims to develop control systems for the fusion power industry after raising $3.7 million in pre-seed funding. The company focuses on addressing the need for specialized components in fusion technology, particularly in magnetic confinement fusion, and plans to create adaptable control systems for various startups in the sector.

Bias: 14 Sentiment: +0.10
Hacker News — Front Page

Challenges in AI Adoption Related to User Discovery

AI adoption faces significant barriers due to users' lack of awareness about its capabilities, referred to as the 'discovery problem.' While templates and contextual suggestions can assist users, the current systems require users to know what to ask, limiting their potential. Enhanced interfaces are needed to better reveal AI capabilities.

Bias: 4 Sentiment: +0.00
New York Post

Iran claims to have captured US submarine drone in Strait of Hormuz

Iran's Revolutionary Guards reported the capture of a US uncrewed submarine drone, identified as a Dive-LD, in the Strait of Hormuz. The Pentagon confirmed that an underwater drone had malfunctioned but stated it was an older model that did not collect sensitive data. Anduril, the manufacturer, acknowledged the loss but emphasized the drone's design for operation in hazardous environments.

Bias: 4 Sentiment: +0.00
BBC — Business

Singaporean Man Pleads Guilty to Cryptocurrency Theft in the US

Malone Lam, a 22-year-old Singaporean man, pleaded guilty to a racketeering conspiracy charge in the US for organizing a cryptocurrency theft totaling $245 million. He admitted to leading a group that deceived victims and laundered the stolen funds, which were used for extravagant purchases including luxury cars and high-end goods.

Bias: 4 Sentiment: -0.10
BBC — World

Report finds Meta ran ads promoting child sexual abuse material in India

A report by the Tech Transparency Project revealed that Meta ran 332 ads featuring child sexual abuse material on its platforms, with 84 ads in India. This follows a government order for Meta to remove such content. The report highlights ongoing concerns about the effectiveness of Meta's ad moderation systems and calls for stricter reporting requirements for social media platforms in India.

Bias: 4 Sentiment: -1.00
Ars Technica

LG TVs Under Scrutiny for Potential User Tracking Capabilities

LG TVs are facing scrutiny after a video suggested they can track users even when powered off. Gamers Nexus, in collaboration with Level1Techs and security researchers, analyzed the tracking capabilities and found that LG TVs were scanning local networks for other devices.

Bias: 4 Sentiment: -0.10