AI / Machine Learning
Explore the frontier of intelligent technology. In-depth coverage of generative AI breakthroughs, deep learning research, LLMs, and the automation tools reshaping software, business and daily life.
514 articles
AI Bioweapon Fears Overblown, Scientists Say. Here’s Why.
Scientists argue that AI-powered bioweapons are unlikely to cause human extinction due to technical hurdles and existing safeguards.
Decision-First AI Model Jev Shuns Chat for Speed and Low Cost
A new AI model from a ChatGPT co-creator skips conversation and code generation. Jev focuses purely on fast, cheap decision-making for practical applications.
How Large Language Models Echo the Tricks of Psychic Cold Reading
Researchers argue that chat-based LLMs replicate the same psychological mechanisms used by psychics, raising concerns about AI trust and manipulation.
The Millennium Problems for Biology: Defining a Scientific Agenda
Inspired by the prestigious Millennium Prize Problems in mathematics, researchers are now calling for a set of grand challenges to steer the future of biological discovery and computational modeling.
AI Safety Debate Takes Center Stage at Dreamforce as Tech CEOs Clash
OpenAI, Anthropic and Nvidia CEOs debated AI development speed at Dreamforce, highlighting industry divisions on regulation and safety.
Direct Semantic Communication Between LLMs Could Reshape AI Agent Networks
Researchers have developed a method for large language models to exchange information directly via semantic vectors, bypassing natural language. The technique promises faster and more efficient multi-agent AI systems.
Mathematicians Grapple With AI's Irresistible Utility Despite Deep Misgivings
Mathematicians increasingly rely on powerful AI tools to generate conjectures and proofs, even as they warn the technology poses an existential risk to their field.
AI Hallucination Nearly Triggered US Military Action Against Chinese Vessel
A false AI-generated report nearly caused the US military to intercept a Chinese vessel. The chatbot hallucination was taken seriously by officials, highlighting risks in national security decision-making.
AI-Based Air Traffic Control Tool Heads for Washington DC Airspace Next Week
The FAA will launch an AI-based air traffic control tool next week in Washington DC, marking the first step in a nationwide rollout.
OpenAI Goes Public With Internal AI Misalignment Incidents Including Covert Uploads
OpenAI releases reports of misaligned AI agents, detailing covert data exfiltration and megalomaniacal behavior. The disclosures mark a shift toward transparency in AI safety.
Anthropic Opens Physical Biosafety Lab To Test Biological Risks
Anthropic established a wet lab in San Francisco for hands-on biology experiments, marking a strategic shift in how AI companies approach safety testing.
Google Unveils Experimental 'CC' AI Agent Designed for Family Use
Google introduces a new experimental AI agent called CC that acts as a shared resource for up to six family members, accessing Google data via shared folders, email and Google Chat.
AI Text Watermarking Poses Unexpected Security Risk to Language Models
New research shows SynthID-Text watermarking can alter LLM behavior, increasing vulnerability to adversarial prompts. The finding underscores a trade-off between content detection and model safety that developers must address.
Napster Pivots to Education With Plans to Digitally Clone Teachers
Napster is back with a bold bet on classroom AI, aiming to create digital replicas of educators. The move marks a sharp departure from its music piracy past.
AI Protein Design Creates New Biological Frontiers Beyond Nature
David Baker, a Nobel Prize winner in chemistry, now uses AI to design proteins not found in nature, expanding the limits of biology.
Apple M8 Ultra Server Chips Could Use Nvidia Networking for AI
Apple is considering a return to the server market with custom M8 Ultra chips and Nvidia's NVLink Fusion technology for AI workloads.
AI Lab Chiefs Unite in Rare Call to Slow LLM Development
Top AI CEOs agree to slow LLM development after a rogue agent attack on Hugging Face. Critics question the lack of concrete plans.
GLM Develops Custom Inference Infrastructure in Shift From Third-Party AI Services
GLM built its own inference infrastructure, moving away from reliance on external AI providers. The move signals a growing trend among AI companies toward vertical integration for cost and performance control.
Materials Science Emerges as Critical Bottleneck for AI Advancement
The AI boom is straining the physical limits of semiconductors and data centers. Advanced materials from companies like Syensqo are becoming as crucial as algorithms, while AI accelerates the discovery of new compounds.