This week's episode dives into a wave of new research shaping how AI is actually being used in education. We explore what works (and what doesn't) when it comes to AI-generated feedback, including why blended, "hybrid" feedback may be the most effective approach - and why more feedback doesn't always lead to better outcomes. The conversation then turns to one of the most important emerging issues: bias in AI systems. From subtle differences in tone to stereotyping based on student characteristics, the research highlights why educators need to be cautious about the data they provide AI tools. "If you use AI to write feedback, it does not treat every student the same way equally." We also talk about the growing evidence around AI tutors - where they outperform humans, where they fall short, and what actually drives meaningful learning gains. Along the way, we tackle major questions around detection, student use, teacher workload, and whether AI can ever replace human connection. The big takeaway? AI is powerful. And how we design, guide, and use it in education matters more than ever.

Research Papers discussed this week

AI for Feedback

Directive, metacognitive, or a blend of both? A comparison of AI-generated feedback types on student engagement, confidence, and outcomes https://doi.org/10.1016/j.caeai.2026.100553
AI assistance in peer feedback provision: Pedagogically sound, but minimally adopted https://www.sciencedirect.com/science/article/pii/S0360131526000291
Marked Pedagogies: Examining Linguistic Biases in Personalized Automated Writing Feedback https://arxiv.org/abs/2603.12471

AI and Bias

The Life Cycle of Large Language Models: A Review of Biases in Education https://bera-journals.onlinelibrary.wiley.com/doi/10.1111/bjet.13505

AI Tutors

AI tutoring can safely and effectively support students: An exploratory RCT in UK classrooms https://arxiv.org/abs/2512.23633v1
LearnMate: Enhancing Online Education with LLM-Powered Personalized Learning Plans and Support https://dl.acm.org/doi/10.1145/3706599.3719857
Effective Personalized AI Tutors via LLM-Guided Reinforcement Learning https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6423358
Unifying AI Tutor Evaluation: An Evaluation Taxonomy for Pedagogical Ability Assessment of LLM-Powered AI Tutors https://arxiv.org/abs/2412.09416v1

AI Detection

Trusting AI to detect AI? A systematic evaluation of the reliability and robustness of current AIGC detection tools for student academic work (paywalled) https://www.sciencedirect.com/science/article/abs/pii/S0360131526000540

Teacher Workload

Shiksha Copilot: Teacher-AI Collaboration for Curating and Customizing Lesson Plans in Low-Resource School https://arxiv.org/pdf/2507.00456v3

Student use

The Secret Life of Students project - WonkHE Feb/March 2026 https://wonkhe.com/wp-content/wonkhe-uploads/2026/03/Wonkhe_SLOS2026_Jim_slides.pdf
Is a random human peer better than a highly supportive chatbot in reducing loneliness over time? https://www.sciencedirect.com/science/article/pii/S0022103126000417?dgcid=rss_sd_all

Inside the latest AI in education research: tutors, bias, and impact