The News, Explained

OpenAI published a report on responsible AI use in education on November 10. Drawing on education partnerships in EMEA, meaning Europe, the Middle East and Africa, it covers educator support, local conditions, AI literacy and evaluation. AI literacy means being able to examine a tool’s answers, understand its limitations and judge when to use it. OpenAI says the early country experiences should be read within their particular education contexts. Source Source

The ThinkYoung student survey cited in the report covered 5,000 secondary students aged 13–19 in ten European countries. Data were collected from April 14 to May 15, 2026, with 500 participants per country. The sample was structured by gender and region, and responses were weighted using population statistics. A sample is the group that actually takes part in a survey. Those countries, ages and respondents define the starting point for reading its figures. Source

Separating the questions makes the guidance gap more concrete. Seventy-nine percent reported having used AI for homework or assignments. Forty-seven percent said teachers had shown them how to use AI for assignments, while 57% agreed their school had clear AI-use guidelines. These items ask about experience, instruction and school rules respectively. Each percentage needs to be read with the question it answers. Source Source

For teachers’ experience, OpenAI’s report describes an Estonian survey from spring 2026. A group of 723 respondents estimated average weekly time savings of 1.8 hours from AI. This is a time estimate teachers reported about their own work. It can inform questions about whether educators gain more room for their tasks. Source

Students’ experience appears in a separate case. Citing the first-year review of Estonia’s ITI learning app, OpenAI’s report gives average student satisfaction of 3.07 out of five, with a sample of 182. It also describes feedback about excessive questioning, slow exchanges and wanting clearer explanations. This indicator concerns students’ experience of that particular app. Its population and question differ from the teacher time estimates. Source

OpenAI distinguishes account activation, usage and satisfaction as signals observable during adoption. Activation concerns starting to use a service, usage measures describe patterns of use, and satisfaction captures an assessment of the experience. Lasting knowledge, independent judgement and improved ability to complete tasks are presented as subjects for sustained research and evaluation. The report advocates measuring use and educational outcomes separately. Source Source

OYOPICK’s Take

The practical question we take from these findings is what guidance should follow the question of how many people use AI. Responses about having school rules differ from responses about being shown how to use a tool. We welcome teachers and students discussing both the permitted scope of help on an assignment and how to examine an answer. Expanding learners’ choices calls for access to tools and the ability to judge their use to develop together.

Feedback about wanting clearer explanations even when an app asks many questions also deserves attention. In judging educational AI’s progress, we hope the volume of questions or usage is considered alongside what a student has understood. One possible activity is to ask learners to explain the same concept in their own words, or solve a new problem independently, after receiving AI assistance. We hope such evaluation helps examine teachers’ capacity and students’ independent learning as distinct goals.