California’s attorney general has subpoenaed OpenAI as part of an investigation into cybersecurity incidents and model risks. The demand for information is not a finding of wrongdoing.
Read articleCalifornia’s attorney general has subpoenaed OpenAI as part of an investigation into cybersecurity incidents and model risks. The demand for information is not a finding of wrongdoing.
Read article
A Stanford-led audit found AI models scored differently through developer APIs and consumer chat interfaces, suggesting benchmark results
MIT researchers used a two-stage AI controller to make an insect-scale flying robot 447% faster, with 255% greater
A study of 609 young adults found AI transparency changed the link between rational reactions to in-feed adverts
Six experiments found people were more willing to behave dishonestly towards AI agents, partly because they expected less
A study of 847 female online shoppers linked AI nudges and social cues with addictive shopping behaviour, while conversational AI showed a surprisingly different pattern.
Read articleA peer-reviewed comparison found AI financial advice showed less projection bias than human advisers but tended to recommend
University of Maryland researchers taught a robot system to learn how objects should move from ordinary videos, then
Nvidia’s new 64GB DGX Spark starts at $4,999 and can link two boxes into a 128GB local AI
Anthropic is putting $100 million behind a programme to train 10,000 engineers who can turn Claude projects into
HardFlow steers flow-matching generative models so their final outputs satisfy formal hard constraints. The method was tested on robotics, physical control and image tasks, but real-world safety still depends on the constraints and surrounding systems.
Read articleCarnegie Mellon researchers developed LAMP, a planning framework that lets teams of robots reason about the object's route
NielsenIQ says 51% of US consumers in its recurring tracker used at least one AI-powered tool to support
Anthropic is interviewing Claude users about what they want from AI and letting participants publish the complete transcript.
Microscopic robots equipped with temperature sensors and artificial cilia can coordinate their movements to pump fluid and reshape
Researchers analysed real ten-minute speed dates using facial, audio and language signals. Their models found suggestive patterns, but the uncertainty intervals did not establish a clear advantage over a random baseline.
Read articleA survey of 360 university students in China found that those who saw generative AI as more empowering
A new US executive order directs major federal online services towards America.gov, a conversational front door for questions
Meta has expanded Muse into a small-business agent that connects with tools such as QuickBooks, Shopify, Slack and
A quantum-optical spin glass stored and recalled far more patterns than a classic associative network in a small
A randomised trial found junior patent lawyers gained most while using AI, yet senior lawyers showed the clearest improvement later without it. Better AI-assisted output does not necessarily mean better long-term learning.
Read articleMIT Sloan researchers found that AI projects can stall when employees are left to absorb testing, review and
Scientists report saving nearly seven hours a week with AI, but faster analysis is creating a backlog of
MIT engineers built a paper-thin swimming robot powered by one layer of living muscle cells. The slow laboratory
The European Commission has opened a targeted copyright consultation covering AI training, licensing and synthetic imitations of performers.
A survey of 11,000 consumers suggests many people grant AI access even while feeling uncomfortable about it. The result exposes a gap between the permission companies can measure and the trust users actually feel.
Read article
Anthropic says Claude Sonnet 5.5 is over 30% faster and can cost less per completed task while keeping
OpenAI says GPT-6.1 Sol delivers capabilities close to GPT-6 Astra with a cheaper, faster operating profile. The bigger
A satellite AI model has mapped smallholder crops in Senegal using far less labelled data than conventional approaches
Microsoft is expanding a Copilot agent called Autopilot that can keep working in the cloud after you leave,
A Korean research team’s quadruped robot completed a 42.195-kilometre marathon in 4 hours 19 minutes without changing its battery. The result shows how better hardware and learned movement can stretch the practical range of legged
Read article
Bristol researchers have shown mathematically that a complex image classifier can be built from a small number of
Citizens Advice says essential-service customers should have a right to reach a person within a reasonable time instead
Google’s first Project Suncatcher orbital test is due to launch four TPU AI chips on 1 October. It
Meta has introduced camera-free Ray-Ban AI glasses focused on audio, calls, translation and its Muse assistant. Removing the
A Rutgers survey found AI use has risen sharply, yet fewer than one in ten respondents wanted unsupervised AI making consequential decisions such as hiring, lending or college admission. Heavy use and public caution are
Read articleSan Francisco Fed researchers found that banks with more AI-related hiring devoted a smaller share of their lending
A Stanford study across 41 countries found that firms adopting generative AI reduced the junior share of employment
Anthropic says a scam network used thousands of AI-generated dating personas across more than 20 apps, with real
Researchers used generative protein models to design new enzyme components, built hundreds of variants and found some synthetic
In a 57-person fMRI experiment, a short pre-lecture conversation with a GPT-4-powered AI instructor improved learning and neural alignment relative to no interaction.
Read article
Synapse combines episodic and semantic memory with spreading activation to help AI agents retrieve the right past information
A small study found children readily trusted a generative-AI toy, even on first encounter, raising a practical question
MorFiC transfers a locomotion policy trained on one quadruped to different four-legged robot bodies without fine-tuning, including real
A peer-reviewed food-design study trained AI on human burger recipes, watched it rediscover a Big Mac-like formula and
Ten bias-audit instruments mostly detected occupational gender bias in frontier AI models but failed to agree on a stable ranking of which models were better.
Read article
The UK and US have launched an AI and autonomy defence partnership focused initially on protecting critical infrastructure
California has signed seven data-centre laws covering water, electricity and local oversight as AI infrastructure expansion meets growing
JudgmentLens helped 16 non-expert readers complete legal-reading tasks faster and feel more confident, but scored comprehension did not
An audit of 1,536 AI product responses found recommendations and displayed source domains could change sharply across repeated