AI Research
The 35 most recent ai research briefs, newest first.
-
November 3, 2026 · AI Research
When Modernization Breaks Working Code: A Study of LLMs on Legacy PHP
What happened A study tested five LLMs on a single file modernization task: each model was given 100 legacy PHP files from WordPress 4.0 and asked to modernize them. The research examines whether LLM...
-
October 30, 2026 · AI Research
Sandboxing LLM-Generated Code Needs a Different Threat Model
What happened LLM generated code introduces a distinct security challenge: failures tend to be accidental rather than adversarial, which means threat models built around malicious actors may not appl...
-
October 30, 2026 · AI Research
Research Insight: Kinsus Tech identifies three challenges for larger AI chip packages
What happened At the 2026 TPCA Show, Kinsus Interconnect Technology R&D vice president Ting Hao Lin presented findings on the packaging challenges created by AI computing demand. Lin identified three...
-
October 28, 2026 · AI Research
AI Images Can Obscure the Reality of Factory Farms
What happened Researchers studying AI generated images of farm animals found that a hidden feature within the image generation system steered outputs toward idealized, pastoral scenes rather than dep...
-
October 26, 2026 · AI Research
A multi-source earth observation and artificial intelligence framework for unsupervised water-use anomaly screening in Mediterranean agriculture
What happened Researchers present AI4WaterGuard, a framework combining multi source earth observation data with unsupervised AI to screen for agricultural water use anomalies across large Mediterrane...
-
October 19, 2026 · AI Research
Claude Traces Five NIST Curve Seeds to 1997 and 1998 Timestamps
What happened A disclosure dated October 16 describes Claude tracing five NIST elliptic curve seeds to an old computer program that used timestamps from 1997 and 1998 as inputs. The Bitcoin Foundatio...
-
October 17, 2026 · AI Research
Using AI to help find Mexico’s legions of missing children
What happened Two AI tools are under development aimed at helping locate missing persons in Mexico, where more than 134,000 people are currently being sought. The effort responds to a large scale dis...
-
October 17, 2026 · AI Research
Conversational brain AI for adaptive epilepsy care
What happened A small study found that patients can help calibrate their own seizure detection systems by conversing with an AI agent. The approach allows individuals to provide feedback that fine tu...
-
October 16, 2026 · AI Research
Spatiotemporal AI-based modeling of gene regulation and variant effects in the developing human brain using Seal
What happened Researchers introduced Seal, an AI framework designed to predict how regulatory genetic variants affect gene expression across 802 distinct brain cell states. The system models spatiote...
-
October 7, 2026 · AI Research
Decoding systemic vascular health and hypertensive disorders in pregnancy through retinal imaging and Visionary AI
What happened Researchers developed an interpretable AI framework called Visionary AI that predicts hypertensive disorders of pregnancy by analyzing retinal images combined with vascular modeling. Th...
-
October 5, 2026 · AI Research
Commentary: AI can be more comforting than a person. Here's what humans can learn from it
What happened Psychology researchers from the universities of Manchester and Durham ran five experiments comparing emotional support messages written by people against those produced by large languag...
-
October 2, 2026 · AI Research
AI Voice Screen Caught 82% of Type 2 Diabetes Cases but Falsely Flagged 47% of People Without It
What happened An AI tool that analyzes roughly 20 seconds of spoken audio correctly identified 82% of type 2 diabetes cases in a large UK study, according to findings reported by Medical Daily. The m...
-
September 30, 2026 · AI Research
Erdős Unit Distance Problem: What It Is and What AI Found
What happened The Erdős unit distance problem, posed in 1946, asks how many pairs of points in a plane can be placed exactly one unit apart. Paul Erdős offered a conjecture that held for roughly eigh...
-
September 28, 2026 · AI Research
Financing of AI build out raises risks in US, researcher says
What happened A researcher claims the AI infrastructure buildout is on track to consume a larger share of U.S. economic output than prior transformative buildouts, including electricity, railroads, i...
-
September 26, 2026 · AI Research
China publishes 'landmark paper' on AI-to-AI technique that kicks human 'bottleneck' out of the loop and replaces us with an AI 'modem' — C2C brainwave direct connection achieves 150% boost in inference speed
What happened Researchers at Tsinghua University published a method called Cache to Cache (C2C) that lets separate AI models exchange internal representations directly, bypassing the step of generati...
-
September 22, 2026 · AI Research
AI co-scientists are revolutionizing how research is done
What happened AI systems described as co scientists are now capable of generating hypotheses, designing experiments, and analyzing data autonomously, according to reporting in Nature. A biochemist id...
-
September 19, 2026 · AI Research
Anthropic says its chatbot Claude is taking over the work of building its own successor
What happened Anthropic reported that Claude now leads 26 percent of the company's research and development work, according to a company statement attributed to a Washington Post report dated Septemb...
-
September 2, 2026 · AI Research
Ensuring multiomics data reproducibility for artificial intelligence with reference materials as a common calibrator
What happened A study published in Nature proposes that reference materials should be adopted as a common calibrator across multiomics measurements. The core recommendation is that reference material...
-
August 26, 2026 · AI Research
Show HN: Poka-Yoke – mistake-proofing skills for Claude Code, with a benchmark
What happened A GitHub project applies Shigeo Shingo's poka yoke manufacturing concept to AI assisted coding via a Claude Code plugin. The tool audits code for error prone patterns, designs APIs to r...
-
August 25, 2026 · AI Research
Can AI ever be conscious? The question stems from a misconception
What happened A book by Anthony Chemero, published by Columbia University Press, argues that AI, psychology, and philosophy have collectively neglected a key factor in understanding consciousness: th...
-
August 21, 2026 · AI Research
When AI explains its decision, humans may stop thinking independently
What happened A new study examined how AI recommender tools affect human judgment when reviewing innovation proposals. Researchers found that AI tools were persuasive enough to shift human decisions,...
-
August 20, 2026 · AI Research
AI’s attribution problem gets worse as models scale
What happened Researchers associated with MIT's Computer Science and Artificial Intelligence Laboratory trained 24 model ensembles on datasets ranging from 256 to over 160,000 images drawn from seven...
-
August 19, 2026 · AI Research
AI inference is getting cheaper, but your agents are getting more expensive
What happened Gartner research finds that LLM token costs are projected to fall by 95% by 2030, yet overall AI inference costs are expected to rise because agentic workflows consume far more tokens p...
-
August 16, 2026 · AI Research
The Agentic AI threat cluster: Seven incidents, three actors, and what they mean for your exposure
What happened Tenable's Research Special Operations team reports tracking a cluster of agentic AI threat activity since late July 2026, involving seven incidents and three distinct actors. The team s...
-
August 14, 2026 · AI Research
How to build an AI-driven digital organism
What happened Researchers writing in Nature outline a framework for building what they call an AI driven digital organism, conceived as a system of integrated multiscale foundation models. The propos...