LLM assistants in decision-making roles (e.g.,
purchasing goods, selecting academic submissions, films, CVs ...) will implicitly favor LLM-based AI agents and LLM-assisted humans over ordinary humans as trade partners and service providers.
#AIagentBiais #DecisionMaking
arxiv.org/pdf/2407.128...

Tired of tools that force users into prompt‑coding instead of real creative flow?
This new paper breaks down 5 paradoxes you must design around to build truly co‑creative AI — not executors, but partners in exploration.
#GenAIUX #UXForAI #AICoCreation #DesignResearch #CreativeAI

human) #metacognition framework, #GenAI #metacognitiveDemand in user workflows, and how genIA systems could support, reduce and manage it to reduce the mental load. #CHI2024 arxiv.org/pdf/2312.10893

a great overview on #LLMs metacognitive capabilities #metacognition www.emergentmind.com/topics/metac...

"infrastructure issues" "implementation issues" & #ContextRot : 3 additional "theories" to account for LLMs #IntelligenceDrift
www.ignorance.ai/p/revisiting...

A rare, data‑driven look at #YouTube public content (nov 2022) beyond recommendations: a vast long tail of unseen, low quality videos hidden beneath the algorithmic spotlight.
#ContentEcosystems #MediaStudies
anyone to adapt/replicate this study with agentic AI?
journalqd.org/article/view...

#GenAI upskilling = #metacognition: notice what the model is doing, check your thinking, and adjust your strategy. von Thienen et al. (2023) frame this as #CreativeMetacognition in design thinking—key for better (and safer) #human‑AI creation.
www.frontiersin.org/journals/psy...

LLM's response traits—rather than #LLMpersona- drift during interactions, especially in emotionally charged contexts, leading to harmful behaviors. Authors propose activation capping along a learned "Assistant trait" axis & prompt steering to stabilize response traits.
arxiv.org/pdf/2601.10387

Chain-of-Thought prompting shows diminishing returns in modern LLMs. Non-reasoning models show mixed results—improved averages but increased variability. However CoT is still relevant for learning and transparency/control. #AIprompting #ChainOfThought #Wharton

GenAI creates a paradox: 20/21 students completed the task (vs 20/31 without AI), but many didn't understand their solutions. Instead of solving existing metacognitive difficulties, GenAI compounds them and introduces new ones (Interruption, Mislead, Progression). #metacognition #learningwithAI

arxiv.org/pdf/2512.19189 Great framework for LLM usage study :"configuration work". Workers must discretize tasks, manage cluttering from prompt engineering overhead, attune practices to task-agnostic systems, and experience desaturation of work texture.

#LLMs #HCI #WorkStudies

Your Simple Guide to Understanding How Humans Connect (and Why AI Can't Fully)
The cognitive, compassionate and contagious facets of empathy: why your boss needs all 3 facets & ChatGPT only has 2 out of 3
#Empathy #Neuroscience #Psychology #Leadership #AIvsHuman
claude.ai/public/artif...

AI transforms peer assessment like in the RIPPLE case (operationalising consensus and trust modelling): boosting reliability even if the transparency and overreliance challenges remain valid
#PeerAssessment #AIinEducation #EdTech #Calibration
link.springer.com/article/10.1...

Nice bridge from SRL (self-regulated learning) theory to practice: HHAIRL frames Human‑AI collaboration where AI scaffolds without short‑circuiting metacognition. FLoRA delivers it
arxiv.org/abs/2507.07362
#SRL #HHAIRL #AIinEducation #LearningAnalytics #HCI #arXiv #metacognition

Training and valuing the brain through neurotechnology codeactsineducation.wordpress.com/2025/03/20/t...

Dans Rematch, Gary Kasparov se bat contre un imaginaire technique, une organisation et finalement contre lui-même. La chronique de Louis de Diesbach sur @synthmedia.fr