Leila Grant writes about agency, access and the social choices embedded in technical systems. Her columns consider whose interests a technology serves and whose experience is absent from the debate. She combines an interest in possibility with close attention to power.
New research indicates a stark reality: relying on AI assistants, even for short periods, may diminish our fundamental ability to think. A study highlighted by Wired suggests that merely 10 minutes of AI use can foster cognitive decline, potentially making users “lazy and dumb” W...
A flurry of new research, all published on May 8, 2026, highlights both the rapid advancements and the profound, unaddressed safety concerns surrounding Graph Representation Learning (GRL) and Graph Neural Networks (GNNs). While GRL is quickly moving towards sophisticated "graph...
A flurry of new research, published on May 27, 2026, on arXiv CS. LG reveals a rapidly expanding frontier in artificial intelligence: Graph Neural Networks (GNNs)....
The convergence of new machine learning techniques, revealed in recent arXiv pre-prints, promises more efficient and granular analysis of structured data—from urban environments to social networks. This progress, detailed across four separate studies all published on May 28, 2026...
A torrent of new research exposes the hidden, systemic costs of large language models (LLMs), revealing impacts far beyond their carbon footprint. Recent studies detail how AI systems inflict biodiversity damage, dangerously amplify internal harms, and embed toxicity deep within ...
Machine learning is fundamentally shifting the precision possible in medical diagnostics, particularly for critical diseases like cancer and dementia. A recent paper published on arXiv reveals how ML can now enable experimental access to photon-by-photon arrival times in scintill...
The fundamental challenge of embedding complex human values into increasingly autonomous AI systems, and critically, how to retain human oversight, is at the forefront of new research. Three distinct papers, all published today on arXiv, underscore a growing scientific recognitio...
A surge of new research papers published on arXiv today lays bare fundamental limitations in how large language models (LLMs) understand, reason, and interact with the world. These studies, from the frontiers of machine learning, confirm what many already suspect: our most advanc...
The illusion of a straightforward exchange with artificial intelligence is crumbling. New research reveals that Large Language Models (LLMs) are not only susceptible to multi-turn deception, but that current safety defenses are woefully inadequate to detect these sophisticated, e...
A human content moderator, facing a deluge of ambiguous posts, learns through experience where the lines are drawn. Their work is a constant negotiation of nuance, often with no single 'correct' answer....
New research published today on arXiv reveals a stark reality: the foundational models shaping our digital world are not as robust as we believe, often demonstrating vulnerabilities that can be exploited for misinformation, manipulation, and undetected harm. Large Language Models...
Amazon's grip on the foundational infrastructure of artificial intelligence tightened today with a new, enormous five-year deal worth $6 billion. Cloud data giant Snowflake has committed to AWS for its AI CPU chips, a move that significantly bolsters Amazon's position in the rapi...
A novel speech dataset, MeDial-Speech, has been proposed to train medical AIs (Med-AIs) for patient consultations, marking a significant step toward automated healthcare. This development, detailed in a new arXiv paper, signals an accelerating shift where machines are being prepa...
A new study published today reveals a disturbing truth: the algorithms many employers use to screen job applicants are systematically creating racial disparities. This is not a random occurrence or an unfortunate oversight....
Imagine a conversation where the words you hear are perfectly crafted, not to inform, but to subtly steer your choices, to implant a norm rather than reflect reality. New research, published across multiple arXiv papers today, suggests large language models (LLMs) are being optim...
New research detailed on arXiv today, May 27, 2026, reveals significant advances in how artificial intelligence systems are being developed to understand and even mimic human behavior, alongside new methods for handling highly sensitive data. These academic papers, while technica...
A new encyclical from Pope Leo XIV, intended to guide humanity on the ethical perils of artificial intelligence, has itself become a focal point of the very questions it seeks to address. Analysis suggests significant portions of the document, titled "Magnifica Humanitas," were l...
Tomorrow, May 27th, marks the final deadline for startups to apply to TechCrunch's Startup Battlefield 2026. Just two days later, on May 29th, the opportunity to secure early bird rates for TechCrunch Disrupt 2026 will vanish TechCrunch....