ai ethics research unesco call for papers ens paris a mortarboard cap with a tassel

Call for Papers: Advancing Research on the Ethics of Artificial Intelligence

September 23, 2026
Artificial Intelligence
Comment off

On 23 September 2026, UNESCO and the AI and Society Institute at ENS-PSL launched a call for papers for researchers who completed a PhD within the past five years. Applicants submit an anonymised 800 to 1,000-word extended abstract by 30 October on any of 14 themes framed by UNESCO’s 2021 Recommendation on the Ethics of Artificial Intelligence. The laureates and jury of the UNESCO Beruniy Prize will review proposals, successful authors will be notified by 15 November, full papers of 5,000 to 7,000 words are due on 1 December, and selected authors present at ENS in Paris on 10 December. This guide covers eligibility, the timeline, the seven required abstract elements, the jury, the publication route and the mistakes that sink submissions.

humanoid robot control narrow gaps whole body ai a portal frame with two uprights and a lintel

Humanoid Robots Navigate Narrow Gaps and Obstacles With Whole-Body AI Control

September 23, 2026
Artificial Intelligence
Comment off

TANGO is a whole-body vision-language-action model from UC Berkeley and Princeton that takes a spoken instruction and a camera feed and outputs motion for all 29 joints of a humanoid, rather than planning a flat two-dimensional path. Trained on roughly 65,000 physics-verified trajectories synthesised in about 211 GPU-hours, it succeeded 53% of the time in simulation against 27% for the 2D baseline, and cut real-robot collisions from about 16% to about 10% using RGB cameras alone. This article covers the architecture, the arithmetic and the stated limits.

ai traffic planning framework cities forecast a milepost with a tapered shaft and a flat cap

A New AI Framework Could Help Cities Plan for Future Traffic

September 23, 2026
Artificial Intelligence
Comment off

Researchers at NYU Tandon have published a geospatial AI framework that pairs ARIMA and LSTM forecasting with H3 hexagonal hotspot analysis and a locally hosted LLaMA query portal, so planners can ask questions of fifteen years of New York traffic data in plain English. The neural model cut root mean square error to 342.56 vehicles per day against ARIMA’s 417.62, and the framework projects average daily volume rising from 12,540 in 2025 to 19,680 in 2029. This article works through the arithmetic, the quotes and the limits the authors state.

ai disagreement study perception not opinion a broad keystone block tapering to a narrow flat top

When AI Disagrees, People Change How They Perceive the Technology, Not Their Original View

September 23, 2026
Artificial Intelligence
Comment off

A University of Michigan study of 482 US adults, published in Computers in Human Behavior Reports, tested what happens when a chatbot pushes back on a user’s reasoning about personal dilemmas. Agreement raised confidence; disagreement did not lower it. What changed instead was how participants rated the AI, which they judged less emotionally capable and more machine-like, and how willing they were to use it again. This article sets out the design, the asymmetry, the limits, and what it means for anyone engineering sycophancy out of a model.

ai gender bias womens language workplace prompts a gramophone with a flared horn on a square box body

AI Might Be Making Women Sound Bad at Work

September 21, 2026
Artificial Intelligence
Comment off

Johns Hopkins researchers took 427 real workplace writing prompts, rewrote each into a women-associated and a men-associated version, and ran both through GPT-4, Llama, Mistral and Gemma. Every model returned shorter, plainer, less formal documents for the women-coded phrasing. Adding a male or female sign-off name changed nothing. Linear probes decode the linguistic register at 0.988 accuracy by layer 5 while name gender reaches only 0.717, and activation patching puts the causal weight in layers 0 to 7. This article covers the method, the per-model results, the two controls that rule out mirroring, and why the authors say users cannot fix it themselves.

physical ai hardware becomes the neural network a physical ai honeycomb cluster of seven hexagonal prisms

Toward Physical AI: When the Hardware Becomes the Neural Network

September 21, 2026
AI Research
Comment off

On 18 September 2026 Nature Reviews Physics published a 165-reference Perspective arguing that self-organising memristive networks can learn physically, with no separation between the model and the hardware running it. Six researchers from Los Alamos, INRiM Turin, UCLA, Politecnico di Torino, the University of Canterbury and the University of Sydney set out what has been measured: 93.4% on MNIST learned online on a nanowire device, reservoir computing that trains only a readout layer, and criticality as a tuning target. This piece separates the measured results from the ambition, sets the energy argument against the IEA’s data-centre projections, and lists the four open problems — variability, drift, scaling and readout — that stand between a review paper and a component you can buy.

object recognition ai worse than humans iscience study a jigsaw puzzle piece lying flat and tilted

AI Still Isn’t as Good at Recognizing Objects as People Are, New Test Shows

September 17, 2026
Artificial Intelligence
Comment off

A study by Mugihiko Kato and Biyu J. He of New York University, published in iScience on 17 September 2026, built an image set that systematically untangles global shape, internal parts and texture, then compared human viewers against more than 200 deep neural networks. No model reproduced the human cue-reliance profile, and every model tested substantially underperformed people when global shape was the only usable cue. This article covers the design, the finding that brain-alignment scores did not predict behavioural alignment, which model families came closest, and what the result means for autonomous driving, robotics and assistive devices.

reverse problem generation computational ai mathematics a pyramid frustum with a flat square top

AI Expands Mathematics: A New Frontier of Reverse Problem Generation for Computational AI, Q&A

September 17, 2026
AI Research
Comment off

Associate Professor Hiroshi Kera of Chiba University builds the answer first and constructs the problem around it — a reversal that turned an unglamorous data-generation chore into two new algebraic theorems. We set out what reverse problem generation is, the published runtimes behind it, how accurate the resulting Transformer models actually are, and why Kera expects a field he provisionally calls “AI Algebra” to grow out of the approach.

crowd navigation compact ai training crowded environments a roundabout island disc with raised centre hub

Compact AI Training Method Reduces Navigation Conflicts in Crowded Environments

September 17, 2026
AI Research
Comment off

A paper accepted to ECCV 2026 by researchers at DGIST and KAIST identifies Skill Conflict, a phenomenon in which motion planning and motion prediction compete for the same weights inside the compact shared encoders that mobile robots run. Their fix, Disjoint Parameter Training with Sparse Merging, cut the JRDB collision rate from 0.0189 to 0.0091 on a 6.44 million parameter model. We read the paper, the supplementary tables and the ablations, including the three-way task split that made results worse.

universal charge injector 2d semiconductors ai chips a four square slabs stacked in a neat tower

One Material, Two Transistor Types: The Universal Charge Injector Pointing Toward Stacked AI Chips

September 16, 2026
AI Hardware
Comment off

A KAIST-led team has reported a single material — degenerately doped tin diselenide — that injects charge efficiently into both n-type and p-type atomically thin semiconductor channels, removing the need for a separate contact metal per polarity. The paper reports a drive current more than a thousand times higher than a nickel electrode in p-type tungsten diselenide, an on/off ratio above a billion in n-type molybdenum disulfide, a subthreshold swing below 70 mV per decade, and a monolayer CMOS inverter with a voltage gain near 340. This article explains the physics, converts the figures into engineering terms, places the result against earlier contact schemes, and sets out the distance between a laboratory inverter and a manufacturable chip.

Call for Papers: Advancing Research on the Ethics of Artificial Intelligence

Humanoid Robots Navigate Narrow Gaps and Obstacles With Whole-Body AI Control

A New AI Framework Could Help Cities Plan for Future Traffic

When AI Disagrees, People Change How They Perceive the Technology, Not Their Original View

AI Might Be Making Women Sound Bad at Work

Toward Physical AI: When the Hardware Becomes the Neural Network

AI Still Isn’t as Good at Recognizing Objects as People Are, New Test Shows

AI Expands Mathematics: A New Frontier of Reverse Problem Generation for Computational AI, Q&A

Compact AI Training Method Reduces Navigation Conflicts in Crowded Environments

One Material, Two Transistor Types: The Universal Charge Injector Pointing Toward Stacked AI Chips

Links

Newsletter

Contact

AI research

Call for Papers: Advancing Research on the Ethics of Artificial Intelligence

Humanoid Robots Navigate Narrow Gaps and Obstacles With Whole-Body AI Control

A New AI Framework Could Help Cities Plan for Future Traffic

When AI Disagrees, People Change How They Perceive the Technology, Not Their Original View

AI Might Be Making Women Sound Bad at Work

Toward Physical AI: When the Hardware Becomes the Neural Network

AI Still Isn’t as Good at Recognizing Objects as People Are, New Test Shows

AI Expands Mathematics: A New Frontier of Reverse Problem Generation for Computational AI, Q&A

Compact AI Training Method Reduces Navigation Conflicts in Crowded Environments

One Material, Two Transistor Types: The Universal Charge Injector Pointing Toward Stacked AI Chips

Tags

Links

Newsletter

Contact