AI in Education

StudentBench compares AI tutoring, expert human tutoring, and no tutoring across 2,383 participants preparing with GRE quantitative and verbal questions. The study reports equivalent overall learning gains, with AI outperforming human tutoring in five of seven domains on average.

OmniEdu is an open family of foundation models for K–12 learning and teaching. Its training design combines subject knowledge with curriculum grounding, learner diagnosis, and pedagogical scaffolding.

TinyTorch is an open curriculum that asks learners to build a working machine-learning framework in pure Python, from tensors to transformers. Its goal is not to replace PyTorch, but to make the abstractions inside mature frameworks easier to understand.

Microsoft has reached an agreement with the American Federation of Teachers and its New York affiliate on enforceable principles for AI in schools. The terms cover student-data training, AI companions, transparency, and human review of high-risk decisions.

Microsoft researchers introduce StudentSim, a framework for building personalized student simulators from sparse learner data. The simulators reproduce individual responses and adapt after tutor guidance, outperforming GPT-5.4 across three educational domains.

Harvard Business School’s HBS Foundry bootcamp is using HeyGen-created AI avatars of instructors to critique practice pitches and simulated board meetings. The experiment points to a shift from AI chatbots toward more guided, role-based learning experiences.

Google is adding interactive visuals, simulations, quizzes, document summaries, and research workflows to Search and Gemini. The updates position both products as a more complete toolkit for students.

A Yale Executive MBA exam dispute has grown into a 13-count federal lawsuit over AI detection, academic discipline, and due process. The case shows how fragile AI-cheating enforcement can be when evidence depends partly on unreliable detectors.

Public libraries in the U.S. are seeing unusual demand for “Avoiding AI” workshops. The trend points less to blanket anti-technology sentiment than to frustration with AI features being pushed into everyday tools by default.

A new arXiv paper introduces Earthquaker-AI, a hybrid framework that combines Lego WeDo2 robotics, a retrieval-augmented conversational assistant, and rubric-based assessment for earthquake preparedness education.

A new arXiv paper proposes an AI-accelerated framework for professional upskilling that spans knowledge acquisition, content creation, review, teaching, and assessment. Its main contribution is treating workforce learning as an integrated production-and-learning pipeline rather than a set of isolated AI tools.

A new arXiv paper examines whether large language models can generate high-order questions across Basque, Spanish, and English. The study finds that models can produce answerable questions, but teachers only recognize about half of them as genuinely high-order.