Homechevron_rightArchives

Archives

Explore Stories by Date

Today

UN rights chief warns AI poses existential risk as handful of men hold unlimited power

UN High Commissioner for Human Rights Volker Türk warned the Human Rights Council that advanced artificial intelligence could pose an existential risk to humanity. He called for cast-iron guarantees on AI safety, international red lines, and independent verification, criticizing how a handful of men at companies like OpenAI and Anthropic hold almost unlimited power over the technology.

10 Sources

AMD unveils Agentic AI chips at IFA 2026 to run 125-billion parameter models locally on Windows

AMD announced Strix Halo and Gorgon Halo chips at IFA 2026, enabling local compute for 125-billion parameter models on Windows. Microsoft's Pavan Davuluri revealed Project Zenith for developers, while AMD showcased the 96-core Threadripper Halo Station—a personal supercomputer designed to shift computing from command-based to agentic AI that understands user goals.

2 Sources

Yesterday

viaim Rise AI Earbuds Turn Meeting Recordings Into Action With $259 Indiegogo Launch

viaim introduced Rise AI earbuds at IFA 2026 that go beyond recording meetings to automatically generate drafted emails and task lists processable by AI agents like ChatGPT and Claude. The device launches September 8 on Indiegogo at $259, significantly below the expected $399 retail price.

2 Sources

Oracle's AI Infrastructure Boom Faces New Risk: Heavy Dependence on OpenAI Partnership

Oracle heads into earnings with explosive AI cloud business growth of 58-64% and a massive $638 billion backlog, but investor scrutiny has shifted from capital expenditure concerns to a new challenge: roughly 50% of its AI future depends on OpenAI. The company's $20 billion funding plan and negative free cash flow add pressure as analysts question whether Oracle can monetize its AI infrastructure investments.

2 Sources

AI adoption accelerates anti-competitive practices in digital markets, warns CCI chief Ravneet Kaur

Competition Commission of India chairperson Ravneet Kaur warned that AI adoption is accelerating anti-competitive practices like self-preferencing, discriminatory pricing, and market manipulation in digital markets at jet speed. The regulator is studying agentic AI systems that act independently without human intervention, urging enterprises to conduct self-audits for competition compliance.

2 Sources

Scientists Develop Ultrafast Magnetic-Field Pulses to Slash Computer Memory Energy Use by 100x

Researchers at the University of Edinburgh have developed a breakthrough method using ultrafast magnetic-field pulses that could reduce memory energy use by up to 100 times. The approach addresses the sustainability challenges of AI infrastructure by moving magnetic memory closer to fundamental thermodynamic limits, potentially transforming how data centers operate.

2 Sources

China Readies Humanoid Robots for Combat as PLA Accelerates Military AI Development

China's defense establishment is fast-tracking humanoid robots from laboratories to battlefields. The People's Liberation Army envisions combat robots clearing buildings alongside troops within 5 to 10 years. With Chinese manufacturers controlling 95% of global humanoid shipments in 2025, the PLA gains unprecedented access to advancing robotics technology for military applications.

6 Sources

Building AI Agents Is Easy—Making Them Work in Enterprise Is the Real Challenge

Deploying AI agents in enterprises reveals harder problems than building them. Companies now tackle testing, data governance, and orchestration as agentic AI systems move from demos to real business decisions. Three firms show how decision intelligence, trusted data, and automation orchestration define enterprise AI success.

2 Sources

GPT-6 Astra: OpenAI's AI Model Excels at Spatial Tasks But Struggles With Writing

OpenAI released GPT-6 Astra in September 2026, marking a shift from answering questions to executing complex tasks autonomously. The AI model achieves a 72.6% success rate on desktop automation benchmarks and rebuilds Manhattan in Unreal Engine. But independent testing reveals it writes worse than its predecessor while carrying OpenAI's first critical cybersecurity risk rating.

4 Sources

Danny Boyle Defends 30 Seconds of Generative AI Use in Netflix's Ink Amid Industry Backlash

Danny Boyle revealed his upcoming Netflix film Ink contains about 30 seconds of generative AI, used to animate old photographs and create CGI mice. The 28 Days Later director defended AI in filmmaking as a useful tool when used respectfully, but the disclosure has sparked debate about AI replacing traditional VFX artists and ethical compensation practices.

2 Sources

AI-Designed Drug Shows 3-4 Year Biological Age Reversal in Clinical Trial

Insilico Medicine's AI-designed drug rentosertib reduced biological age markers by 3-4 years in a 42-patient trial, as measured by six independent proteomic aging clocks. Published in Nature Biotechnology, this marks the first clinical evaluation of a drug featuring both an AI-discovered target and AI-designed molecule for aging and disease.

4 Sources

Blue Machines AI Launches Aurora Speech-to-Text Model for India's BFSI Sector with Low Error Rates

Blue Machines AI launched Aurora, a multilingual speech-to-text model designed specifically for India's banking, financial services, and insurance sector. The model achieves 1.51% error rate for English and 2.43% for Hindi BFSI conversations while handling code-mixed languages and low-bandwidth telephone connections.

2 Sources

GPT-6 Astra Autonomously Completes Portal Game in 24 Hours, Costs $571 in API Calls

OpenAI's GPT-6 Astra successfully completed Valve's Portal 3D puzzle game autonomously in 24 hours, processing 3,336 tool calls at a cost of $571.18 in tokens. The achievement demonstrates significant progress in AI gameplay capabilities, though the model required pauses between moves to process game states and plan solutions.

2 Sources

UBS Mandates AI Skills for All 2027 Graduate Hires in Global Banking and Markets Division

UBS becomes first major bank to require AI proficiency for all 2027 graduate and intern applicants. Candidates must demonstrate practical AI use to improve efficiency, not just familiarity with tools like ChatGPT. The move signals a fundamental shift in what entry-level banking roles demand as automation reshapes junior banker responsibilities.

3 Sources

Sophos Integrates OpenAI GPT to Verify Real-World Exploit Paths in Managed Risk

Sophos unveiled Exploit Path Verification, an AI-powered capability built with OpenAI GPT cyber models for Sophos Managed Risk. The tool helps security teams validate which vulnerabilities attackers can actually reach in their environment, addressing the challenge faced by over 625,000 organizations that struggle to prioritize thousands of vulnerability findings.

2 Sources

Get all updates into your inbox 💌, Stay Up-to-date for Free 🥳

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Follow topics that matter to you and stay ahead.

Explore

News Categories

Technology Business Policy Startups Health Science Entertainment

Terms Privacy Content Contact Us