One place for everything you can read or buy: free publications, premium knowledge products, insights and curated reading paths.
Two transformation products can be bought outright, without membership. Each has a free preview.
One-time products27 of 27
Coursera has announced the appointment of Tom Savage as its new Chief Legal Officer and Secretary, effective September 21, 2026. This appointment comes at a critical juncture as the company seeks to scale its global operations and integrate its recent combination with Udemy, necessitating robust legal leadership.
Higgsfield AI has reportedly leveraged OpenAI's GPT-6 Astra to significantly accelerate the development and deployment of new video advertising features, enabling faster market entry and enhanced creative tools for small businesses. This demonstrates the potential for advanced AI models to streamline operational processes and foster rapid innovation.
The Berkman Klein Center has discussed the concept of 'AI constitutions' as a mechanism to safeguard humanity by ensuring artificial intelligence models adhere to principles of being honest, helpful, and harmless. This initiative is noted as an ongoing development.
This analysis addresses the concept of the Value of Statistical Life (VSL), a metric used by governments and institutions to quantify the societal benefit of preventing fatalities. VSL is not a measure of an individual's life worth but represents the collective willingness-to-pay for marginal reductions in mortality risk across a population. It serves as a critical tool in cost-benefit analyses for policy decisions related to health, safety, and environmental regulations.
The UK Department for Education has announced a new 'whole-school approach' aimed at providing support to educational institutions for addressing misogyny and counteracting the influence of individuals like Andrew Tate. This initiative signifies a governmental recognition of the issue within school environments.
Learning Commons, an initiative supported by the Chan Zuckerberg Initiative (CZI), has launched an open platform designed to assist educators and developers in creating and assessing artificial intelligence tools for K-12 education. The platform facilitates the use of state standards, curricula, and various datasets to support the development and evaluation of these AI applications.
OpenAI Research has introduced MentalHealthBench, a new benchmark designed to evaluate the helpfulness and safety of AI responses in realistic mental health conversations. This initiative aims to provide an expert-informed standard for assessing AI performance in a sensitive and critical application domain.
Ofsted in the United Kingdom has published guidance clarifying the registration requirements for various childcare providers, including childminders, nannies, and nurseries. The guidance details when registration is necessary, outlines the registration process, specifies associated fees, and lists essential requirements for compliance.
This systematic review explores the application and impact of active, student-centred pedagogical approaches in Physical Education (PE) across school and higher-education contexts, including Physical Education Teacher Education (PETE). It specifically examines pedagogical models, learner and professional outcomes, and implementation conditions from 2005 to 2025. The research methodology involved comprehensive searches across multiple academic databases, with independent review and appraisal of studies.
The provided document is a guide from Ofsted, detailing the Inspection Data Summary Report (IDSR) for primary and secondary educational institutions, including those with a sixth form. Its primary purpose is to offer an overview and guidance on interpreting the data contained within these reports.
Ofsted has announced updates to its Inspection Data Summary Report (IDSR) releases, which will include new functionality, features, and bug fixes. These updates are intended to refine the IDSR reporting mechanism.
OpenAI has introduced the Australian Youth Safety Blueprint, a strategic framework designed to enhance the safety of AI experiences for young individuals. This blueprint consists of six pillars aimed at both protecting and empowering youth in their interactions with artificial intelligence.
Access to academic support is a significant factor in student success, but its availability and utilization are unevenly distributed among students due to factors such as anxiety, fear of judgment, and lack of confidence. General-purpose generative AI tools are often used but can be inaccurate or lack context. This research introduces and evaluates 'Beacon,' a course-specific Retrieval-Augmented Generation (RAG) system designed to offer private, immediate, and module-aligned academic assistance, aiming to mitigate these disparities.
The AI-GRACE framework is proposed to operationalize agentic artificial intelligence (AI) deployments, linking organizational governance with technical implementation. It addresses the critical need for organizations to establish clear validation, control, and observation mechanisms for AI use cases, ensuring they meet intended outcomes and comply with obligations beyond merely assessing model trustworthiness. The framework is derived from professional observations and a synthesis of existing standards and literature.
A recent research study from arXiv investigates the impact of generative AI, specifically ChatGPT-4.5, on undergraduate computer science students' programming performance and learning retention. The study found that students using ChatGPT achieved significantly higher coding performance but exhibited lower immediate and delayed recall of the material compared to a control group using conventional web search. This suggests a trade-off between immediate task success and long-term knowledge retention when leveraging generative AI in technical domains.
Recent research introduces 'Moral Entropy', a Bayesian framework designed to model and learn from disagreement in moral judgment annotations, rather than dismissing it as noise. This framework distinguishes between irreducible disagreement (aleatoric uncertainty) and uncertainty due to insufficient or noisy annotation (epistemic uncertainty). It also enables the auditing of heuristic consensus rules against a calibrated ground truth using entropy methods.
Existing smart-city mobility platforms utilize simplistic 'transit desert' labels that often fail to differentiate between varying user problems in public transit. This research introduces a new diagnostic tool that provides a pathway-based, data-driven approach to distinguish between localized service mismatches and basic service failures. It identifies specific service attributes contributing to these classifications, offering a more nuanced understanding for user-centered service design.
Independent researchers face significant constraints in conducting direct experiments on online social media platforms to gather causal evidence. Traditional workarounds involve trade-offs between experimental properties. The emergence of open social media platforms introduces a new methodology, 'Open Platform Field Experiments' (OPFEs), which offers a distinct approach to enable direct experimentation by independent researchers.
Research explores using Large Language Models (LLMs) to create an automated self-explanation tutor, ESSE, for introductory programming (CS1) worked examples. This tutor aims to transform passive learning into active study by providing immediate feedback on student-generated explanations, addressing the scalability challenge of human assessment in self-explanation techniques.
A recent study investigated the impact of a brief AI literacy intervention on high school students' reliance on Large Language Model (LLM) recommendations. The findings indicate that while widespread over-reliance on incorrect LLM advice was observed (52.1%), the intervention did not significantly reduce this over-reliance. Instead, the intervention led to an increase in under-reliance on LLM recommendations, suggesting a complex and potentially counter-intuitive effect of basic AI literacy efforts.
Enterprises in regulated sectors face significant challenges in securely onboarding Software-as-a-Service (SaaS) platforms, particularly due to the complex interplay of various control domains like Third-Party Risk Management (TPRM), cybersecurity, Identity and Access Management (IAM), and disaster recovery (DR). The current siloed approach to these domains leads to inefficiencies, delays, and increased risk. A proposed control-driven, end-to-end framework aims to integrate these critical areas into a unified SaaS onboarding lifecycle.
A study conducted in southern Peru investigated the impact of a six-week gamified digital mediation intervention on the written American English phonetic competence of 100 first-year secondary students. The research addressed a gap in controlled evidence regarding gamified tools' effects on phonetic knowledge among Latin American secondary students, specifically focusing on written phonetic recognition rather than oral pronunciation.
A study conducted in the Belgrade metropolitan area explored the perspectives of 797 in-service and pre-service educators regarding the integration of generative artificial intelligence (GAI) in education. While a significant portion (39%) reported using GAI tools, a majority indicated no use of specific GAI tools in their educational practices. The research examined perceived contributions, ethical considerations, and professional readiness concerning GAI.
Research from arXiv explores the use of Large Language Models (LLMs) to address the limited access to professional mental healthcare and the high volume of unanswered peer support queries on online platforms. The study introduces the COmmunity-centered Peer Engaged Support (COPES) dataset and a three-axis evaluation framework to assess LLM alignment with community-specific, lived-experience perspectives in mental health support.
Research from arXiv highlights that the appraisal of AI-mediated financial advice by individuals is significantly influenced by the advice source's style, labels, and the specific decision context. An experiment involving U.S. adults across various financial decisions demonstrated that advice style primarily shapes message and safety perceptions, while expert labels specifically enhance perceived source knowledge. Decision context is crucial in influencing risk and safety assessments, with these appraisals strongly correlating with overall advice quality judgments.
Artificial Intelligence (AI) is significantly transforming mathematics education through intelligent tutoring systems, adaptive learning platforms, and AI learning companions. A recent review highlights that these technologies offer personalized instruction, real-time feedback, enhanced learner engagement, improved problem-solving skills, and support for teacher decision-making. The analysis, grounded in connectivism and adaptive instruction, identifies both opportunities for educational advancement and challenges in implementation.
A narrative review of research published between 2015 and early 2026 indicates that Design Thinking (DT) is a promising human-centered, iterative approach for fostering creative thinking in students across diverse educational settings. The study, which analyzed 58 peer-reviewed publications, suggests DT's effectiveness in cultivating creativity through stages such as empathy, problem framing, ideation, prototyping, and testing, addressing a widely acknowledged challenge in education regarding the cultivation and assessment of creative skills.
Premium structure
A premium publication is not a longer briefing: it moves from situation to options, to recommended instruments, to measurement.
01
Situation
Where does the institution actually stand, on evidence?
02
Options
What courses of action exist, and what does each cost?
03
Recommended instruments
What is recommended, in what order, and with which artefacts?
04
Measurement
How is progress evidenced, reviewed and corrected?