Microsoft publishes its third annual AI in Education report

24 June 2026 | Microsoft (survey by PSB Insights)

​

Microsoft’s 33-page report presents the third wave of its annual survey. PSB Insights fielded it between 13 February and 9 March 2026 among 3,345 educators, education leaders (academic and IT) and students aged 13 and over in the United States, the United Kingdom, Australia, Brazil, Japan and Saudi Arabia, and global figures are weighted by World Bank estimates of education spending. Use is widespread: 92 per cent of students, 92 per cent of education leaders and 88 per cent of educators have used AI for school-related purposes at least once, and 58 per cent of leaders say their schools are implementing or scaling AI. Between 65 and 78 per cent of each group say their use has increased over the past year. Daily use is lower, at 53 per cent of leaders, 33 per cent of educators and 27 per cent of students. Students most often use AI to summarise information (45 per cent), to get answers faster (39 per cent) and to start or brainstorm assignments (38 per cent). Educators most often use it to brainstorm lesson ideas (40 per cent) and to create or update lesson plans (37 per cent). Academic integrity is the leading concern for students (41 per cent worry about being accused of cheating) and educators (42 per cent worry about increased plagiarism), while leaders name privacy and security most often (23 per cent). Confidence is high (77 per cent of students, 74 per cent of educators and 91 per cent of leaders), but optimism has declined, falling 21 points among US students and six and seven points among education leaders and educators. The report identifies a guidance and training gap. Four-fifths of leaders rate their institution’s AI guidance as clear, while half of students and teachers describe it as neutral or not provided. Seven in ten leaders say at least half of users at their institution have had AI training, but 77 per cent of students and 53 per cent of educators say they have had none, and 66 per cent of educators and 52 per cent of students want training monthly or quarterly. Fewer than half of educators describe AI literacy as an essential part of basic education for every student. A US spotlight reports that 36 states (plus Puerto Rico) have published official AI guidance for schools, 26 clarify the importance of AI literacy, six have AI literacy in their standards and three have it in graduation requirements. The report closes with sections on preparing students for AI-related work and with recommendations for leaders and educators. Its case studies include Broward County Public Schools in Florida, which reports deploying 20,000 Microsoft 365 Copilot licences and educators reclaiming 6 to 7 hours a week.

What you need to know: This is a vendor report that also promotes Microsoft 365 Copilot and related products. The case studies and recommendations should be read in that light, and the Broward County figures are reported by Microsoft and the district, not independently verified. The most transferable finding is the gap between leaders’ and staff’s accounts of guidance and training: leaders and the people they serve report different experiences of the same institutions. Microsoft interprets the fall in confidence and optimism as people learning how much there is to know, and the survey data do not test that explanation. The sample spans six countries and mixes K–12 with higher education, so conclusions for a single country or sector need care.

Original link: https://news.microsoft.com/source/2026/06/24/microsofts-new-ai-in-education-report-highlights-widespread-adoption-and-increasing-demand-for-support/

EEF trial finds Oak National Academy’s Aila reduced primary teachers’ lesson planning time by 24 per cent

6 October 2026 | Education Endowment Foundation (evaluated by NFER)

The Education Endowment Foundation (EEF) published a randomised controlled trial of Aila, the AI lesson planning assistant from Oak National Academy, evaluated by the National Foundation for Educational Research (NFER). The trial involved 464 teachers in 108 schools. Teachers of Years 3 to 6 who were allocated to Aila spent 24 per cent less time on lesson planning, with no reduction in the quality of their resources, which an independent panel of teachers assessed. Coverage of the report puts the saving at about 49 minutes a week (around 2.5 hours against 3.3 hours for the comparison group) over ten weeks in autumn 2025, and states that the EEF gives the finding a high security rating. Views on the tool were mixed. Many teachers used Aila alongside their existing planning processes, some preferred other tools, and use declined over the trial. Secondary coverage reports that 81 per cent of teachers in the Aila group used it at least once, that it was used for around 7 per cent of lessons, and that fewer than one in five felt it had meaningfully changed their teaching or planning. Teachers found it most useful for unfamiliar topics. The trial measured teacher workload and resource quality and did not measure pupil outcomes. It follows an earlier EEF trial, published in December 2024, in which teachers using ChatGPT saved 31 per cent of their preparation time.

What you need to know: This is a second independent trial to find workload savings from AI lesson planning. The gap between the measured saving and the level of use is the practical lesson: a free tool does not guarantee sustained use. Coverage attributes the low use partly to the tool being designed to plan lessons from scratch, when teachers more often adapted existing materials. Because pupil outcomes were not measured, the result speaks to teacher time and not to learning. Some of the detailed figures above come from secondary coverage of the EEF report.

Original link: https://educationendowmentfoundation.org.uk/projects-and-evaluation/projects/aila-teacher-choices-trial

Teacher Tapp compares teachers’ 2023 predictions of AI use with reported use in 2026

6 October 2026 | Teacher Tapp

Teacher Tapp compared what teachers in England said in April 2023 that they could see themselves using AI for with what they reported doing in September 2026. In 2023, 5,254 teachers answered the question “Could you see yourself using AI tech to help with any of the following?” on 3 April. In 2026, 8,837 teachers answered “In the last month, have you used AI to do any of the following tasks?” on 30 September. Teacher Tapp weights both samples to reflect teacher and school demographics, and notes that the questions and samples differ, so the changes are indicative and not like-for-like. Use for lesson planning and parent communications exceeded the 2023 predictions. In primary, 55 per cent of classroom teachers reported using AI for planning against 39 per cent who predicted it, and the share of senior leaders using it for parent communications rose from about 28 per cent to 56 per cent. Secondary teachers and leaders exceeded their planning predictions by five points, and secondary senior leaders exceeded theirs on parent communications by ten. Uses involving assessment and data fell well short. About 7 per cent of primary teachers reported using AI to analyse attainment data against 34 per cent predicted, and the shortfall for secondary teachers was 25 points. Analysis of behaviour data was 26 points below prediction among secondary leaders and 24 points below among secondary teachers. Marking in secondary was down 24 points for both teachers and leaders, from about a third to under 10 per cent (7 per cent of secondary teachers against 31 per cent predicted). Homework use was below prediction in both phases. Teacher Tapp notes that AI capabilities have changed greatly since 2023 and that some of the predictions may yet prove correct.

What you need to know: The pattern suggests that teachers have adopted AI first for preparation and communication, and less for assessment and data tasks that involve judgement or sensitive information. That fits the EEF finding on Aila above. Because the comparison sets an expectation question against a reported-use question, which Teacher Tapp flags itself, the gaps are best read as direction of travel and not as precise effects. The rest of the newsletter issue covers topics outside AI: lesson disruption by year group and classroom temperature, how senior leaders explain pupils behaving better for them, lesson planning time, Reception Baseline Assessment problems, teachers’ views on the age of criminal responsibility, and staff supervision outside lessons. The planning-time section is relevant to workload discussions. The most common answer was 25 per cent of lesson time (15 minutes for a 60-minute lesson), and early career teachers tended to spend more time planning, with 43 per cent reporting the 25 per cent figure against 50 per cent of other teachers.

Original link: https://teachertapp.substack.com/p/ai-predictions-vs-reality

Common Sense Media research update: 2026 Census, Wilton Park report and digital citizenship evaluation

27 October 2026 | Common Sense Media

Common Sense Media’s 27 October update covers four items. The first is the ChatGPT for Teens risk assessment, covered in a separate entry in this digest. The second is the 2026 Common Sense Census of tech and media use by tweens and teens, published on 27 October as the organisation’s first broad survey of youth media and technology use in five years. In the survey of US children aged 9 to 17, the typical teenager spent over eight hours a day on screens for entertainment and the typical tween five hours, and 57 per cent of children had their own phone by age 11. Teenagers spent about an hour a day using AI, and about half of tweens reported using AI, although Common Sense Media says those platforms are not designed for children that young. Among teenagers, 77 per cent use AI in their free time, but only 11 per cent of those users say they enjoy it a lot. The third item is the report of a Wilton Park policy dialogue, Childhood in the age of Artificial Intelligence, convened in April with the NSPCC and the UK Department for Science, Innovation and Technology and supported by Common Sense Media, in which young people co-designed and opened the discussion. The fourth is an evaluation of Common Sense’s free Digital Citizenship Curriculum by the University of New Hampshire’s Crimes Against Children Research Center. According to the update, 3,908 US students aged 11 to 14 gained about three times as much knowledge as students in control schools, students aged 12 to 14 also reported acting differently online, including protecting their data, and students with moderate technology use showed stronger digital skills than those with low or high use.

What you need to know: The Census gives recent US benchmarks on tween and teen device and AI use. It comes from a survey of US children, so it does not transfer directly to UK or Gulf settings. The Wilton Park report is relevant to UK policy because a UK government department and a UK children’s charity convened the dialogue. The curriculum findings are reported by the organisation that produces the curriculum, and the evaluation report itself was not reviewed for this digest.

Original links: https://www.commonsensemedia.org/research/the-2026-common-sense-media-census-tech-and-media-use-by-tweens-and-teens

https://www.wiltonpark.org.uk/reports/children-in-the-age-of-artificial-intelligence/