{"id":424,"date":"2026-10-31T18:04:35","date_gmt":"2026-10-31T18:04:35","guid":{"rendered":"https:\/\/prowad.pro\/?p=424"},"modified":"2026-10-31T18:04:35","modified_gmt":"2026-10-31T18:04:35","slug":"teacher-ai-evaluation-skills-world-bank-trial","status":"publish","type":"post","link":"https:\/\/prowad.pro\/fr\/teacher-ai-evaluation-skills-world-bank-trial\/","title":{"rendered":"Teacher AI Evaluation Skills: What a 1,269-Teacher Trial Found"},"content":{"rendered":"<p><strong>Giving teachers access to a generative AI tool does not automatically make their working day shorter\u2014or make their teaching better.<\/strong> A randomized study of 1,269 sixth-grade teachers in Lima, Peru, offers a more useful way to think about AI professional development: training may matter most when it improves how educators instruct, question and evaluate the system.<\/p>\n<p>The World Bank working paper <a href=\"https:\/\/documents.worldbank.org\/en\/publication\/documents-reports\/documentdetail\/099309210012625906\" target=\"_blank\" rel=\"noopener\">Not Faster, but Better: Generative AI and Teacher Skill in Peru<\/a>, released on 1 October 2026, reports a striking combination of findings. A package of training, access to Microsoft Copilot and follow-up support increased adoption and improved teachers\u2019 prompts and judgment of AI-generated material. Yet the researchers found no measurable time savings across seven core teaching activities.<\/p>\n<p>For school leaders, teacher educators and training providers, the implication is not that AI training failed. It is that success should be measured through professional judgment and the quality of human\u2013AI work\u2014not simply through minutes saved or the number of teachers who open an account.<\/p>\n<h2>What the teacher AI training trial actually tested<\/h2>\n<p>The study covered 1,269 sixth-grade teachers in 390 public schools in Metropolitan Lima. Schools were assigned to treatment or control groups: 196 treatment schools included 660 teachers, while 194 control schools included 609 teachers. The researchers used a school-level randomized design and followed participants over approximately eight months.<\/p>\n<p>The intervention was a package rather than a single workshop. Teachers in treatment schools received:<\/p>\n<ul>\n<li>A one-day, in-person workshop introducing how large language models work, practical classroom applications and common limitations.<\/li>\n<li>Access to a Microsoft Copilot licence through institutional education accounts.<\/li>\n<li>Weekly WhatsApp follow-up messages with examples, reminders and suggestions for applying the training to real teaching tasks.<\/li>\n<\/ul>\n<p>The workshop was co-designed with 12 teachers and connected AI use to the Peruvian national curriculum. Its prompting routine asked teachers to specify four elements: <strong>who<\/strong> the material was for, <strong>what<\/strong> the AI should produce, <strong>how<\/strong> the output should be structured, and how it should align with the <strong>CNEB<\/strong> curriculum. The training also emphasized that generated lesson plans and worksheets were drafts to be checked, not finished products to accept automatically.<\/p>\n<p>Importantly, teachers in control schools were not prohibited from using AI; they retained access to publicly available tools. The trial therefore tested the added effect of a structured support package, not AI access versus no possible access.<\/p>\n<h2>What improved\u2014and what did not<\/h2>\n<p>The results are more informative when adoption, work time and skill are treated as separate outcomes.<\/p>\n<table>\n<thead>\n<tr>\n<th>Outcome<\/th>\n<th>Reported result<\/th>\n<th>What it means<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Copilot licence activation<\/td>\n<td>+20.4 percentage points<\/td>\n<td>The programme increased sustained access and uptake.<\/td>\n<\/tr>\n<tr>\n<td>Active platform use<\/td>\n<td>+3.15 active weeks over 28 weeks<\/td>\n<td>Teachers used the supplied tool more often.<\/td>\n<\/tr>\n<tr>\n<td>Prompt quality<\/td>\n<td>+0.293 points on the training-based rubric; +0.375 on an independent rubric<\/td>\n<td>Teachers wrote clearer, more specific and better-contextualized instructions.<\/td>\n<\/tr>\n<tr>\n<td>Judging weak AI outputs<\/td>\n<td>9.9 percentage points less likely to misjudge prompt quality; 6.7 points less likely to rate an inadequate response as good<\/td>\n<td>Training strengthened critical review, not just tool operation.<\/td>\n<\/tr>\n<tr>\n<td>Time spent on seven teaching tasks<\/td>\n<td>No statistically significant time-saving effect<\/td>\n<td>Greater adoption did not translate into a shorter reported workday.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>The prompt-quality results are particularly important. The researchers used one rubric based on the workshop and another designed independently of its materials. The estimated improvements were 0.293 and 0.375 points respectively; these represented increases of 9.2% and 11.1% relative to the control means. Improvement on the independent rubric makes it less likely that teachers merely memorized the workshop\u2019s preferred template.<\/p>\n<p>The time-use analysis covered lesson planning, differentiating materials, designing assessments, grading, collaboration, administration and communication with parents. The study did not find significant reductions in time spent on these activities. That does not prove AI can never save teachers time. It means this particular package, in this setting and over this observation period, did not produce measurable savings on those reported tasks.<\/p>\n<h2>Why teacher AI evaluation skills matter more than prompt tricks<\/h2>\n<p>A prompt can be well written and still produce a weak lesson, an inaccurate explanation or a worksheet that does not match the students\u2019 level. Prompt construction is only the first half of competent use. The second half is deciding whether the response is fit for its educational purpose.<\/p>\n<p>The Peru trial assessed this distinction directly. Teachers were asked to review an AI-generated worksheet for a sixth-grade mathematics lesson on multiplying decimals. They had to judge whether it was ready to use, identify deficiencies and explain how they would improve it. Trained teachers were more likely to recognize that the output needed improvement and less likely to accept an inadequate response as good.<\/p>\n<p>This is a deeper skill than knowing which words to place in a prompt. It involves subject knowledge, curriculum awareness, knowledge of learners and the confidence to reject a fluent but unsuitable answer. A teacher who can write a detailed instruction but cannot detect a mathematical error or a developmental mismatch has not yet demonstrated professional AI competence.<\/p>\n<p>A 2026 critical review in <em>Discover Education<\/em> similarly argues that teacher development should include critical evaluation and ethical reasoning as core competencies, not optional additions to technical familiarity. Its synthesis of 20 studies identifies persistent gaps in prompt training, evaluation practice and ethical integration. Read it alongside Prowad\u2019s guide to <a href=\"https:\/\/prowad.pro\/fr\/ai-teacher-training-competency-frameworks-2026\/\">teacher AI competency frameworks<\/a>: the framework article maps what educators may need to know, while the Peru trial provides experimental evidence about one way to develop and measure selected skills.<\/p>\n<h2>Which teachers benefited\u2014and what that suggests about equity<\/h2>\n<p>The researchers found that gains in prompt quality were larger among teachers above the sample\u2019s median age of 54. These teachers had started with lower scores in the control group, and the intervention narrowed rather than widened the age-related gap in AI prompting skills.<\/p>\n<p>This finding challenges the assumption that older educators will necessarily benefit less from AI professional development. It does not mean age alone determines who needs training, or that the same pattern will appear in every school system. It does suggest that programmes should not reserve meaningful support for teachers who already describe themselves as confident technology users.<\/p>\n<p>Equitable professional development should offer multiple entry points: basic explanations for newcomers, subject-specific practice for experienced educators, accessible materials, time to experiment and opportunities to ask questions without embarrassment. The strongest design starts from what teachers can currently do, not from stereotypes about who is \u201cgood with technology.\u201d<\/p>\n<h2>What the study cannot prove<\/h2>\n<p>The findings are promising, but several boundaries matter before a school system copies the programme.<\/p>\n<ul>\n<li><strong>It is a working paper, not a peer-reviewed journal article.<\/strong> The World Bank Policy Research Working Paper Series circulates work in progress to encourage discussion. The paper has a verified reproducibility package, but that is not the same as independent peer review of the claims.<\/li>\n<li><strong>It measures teacher AI fluency, not student achievement.<\/strong> The authors explicitly state that higher-quality prompts do not necessarily produce better classroom instruction or learning. Student learning effects remain an open question.<\/li>\n<li><strong>The intervention was bundled.<\/strong> The trial tested a workshop, a licence and follow-up together. It cannot isolate how much of the effect came from each component.<\/li>\n<li><strong>The setting was specific.<\/strong> The participants were sixth-grade teachers in public schools in metropolitan Lima. Results may differ by curriculum, language, infrastructure, subject and institutional support elsewhere.<\/li>\n<li><strong>Some outcomes were measured through surveys and task scenarios.<\/strong> Platform records offered an objective measure of use, while prompt quality and output judgment were assessed through endline tasks. These measures are valuable, but they do not replace long-term classroom observation.<\/li>\n<\/ul>\n<p>The World Bank\u2019s <a href=\"https:\/\/reproducibility.worldbank.org\/catalog\/608\" target=\"_blank\" rel=\"noopener\">reproducibility record<\/a> describes the trial data and analysis package. The underlying teacher-level dataset remains under embargo pending publication in a peer-reviewed journal, so outside researchers cannot yet independently reanalyse all participant-level records.<\/p>\n<h2>Five competencies to build into teacher AI professional development<\/h2>\n<p>For schools planning their own programme, the trial points toward a practical competency model. The following five areas are an implementation proposal informed by the study; they are not a validated scale from the researchers.<\/p>\n<h3>1. Context-rich task design<\/h3>\n<p>Teachers should be able to state the learner age or grade, prior knowledge, learning objective, curriculum expectation, task constraints and intended format. A useful prompt makes the educational context visible instead of asking for a generic \u201clesson plan.\u201d<\/p>\n<h3>2. Subject and curriculum verification<\/h3>\n<p>Participants should check facts, worked examples, terminology, reading level and curriculum alignment against trusted sources and their own subject knowledge. For mathematics, that means independently solving or checking generated examples; for language learning, it means checking grammar, meaning and level appropriateness.<\/p>\n<h3>3. Pedagogical diagnosis<\/h3>\n<p>Teachers should ask whether an AI-generated activity actually elicits the intended thinking. Does it invite explanation or only answer copying? Does a worksheet scaffold a misconception or reinforce it? Does the material support the intended learners, including those who need accessibility or language adaptations?<\/p>\n<h3>4. Revision and comparison<\/h3>\n<p>Rather than accepting the first response, teachers should practise identifying a specific weakness, requesting a targeted revision and comparing the new version with the original. They should also know when to edit the material themselves or discard it altogether.<\/p>\n<h3>5. Ethical and professional judgment<\/h3>\n<p>Training should address student privacy, bias, copyright, transparency, institutional policy and human responsibility. Teachers need clear rules about what information must never be entered into an external system and which decisions must remain under professional control.<\/p>\n<h2>A four-week pilot schools can adapt<\/h2>\n<p>A school does not need to begin with a large certification programme. It can test a small, measurable cycle using one subject team and one real teaching task.<\/p>\n<ol>\n<li><strong>Week 1 \u2014 Baseline and safe foundations:<\/strong> Ask teachers to complete a short task using an approved tool. Collect the prompt, output and teacher\u2019s review. Introduce privacy rules, common failure modes and the difference between a plausible answer and a verified one.<\/li>\n<li><strong>Week 2 \u2014 Design for a real lesson:<\/strong> Each teacher creates a prompt for an upcoming lesson, specifying learners, objective, format and curriculum link. Colleagues review whether the context is sufficient and whether the output meets the stated goal.<\/li>\n<li><strong>Week 3 \u2014 Critique before use:<\/strong> Give teachers several AI-generated examples containing factual, pedagogical, accessibility or curriculum problems. Ask them to identify the issue, explain its consequence and revise or reject the material.<\/li>\n<li><strong>Week 4 \u2014 Demonstrate transfer:<\/strong> Give participants a new task they did not practise in the workshop. Score their prompt and evaluation independently, then discuss what should change in the next training cycle.<\/li>\n<\/ol>\n<p>For a broader institutional process, connect this pilot to Prowad\u2019s practical guide on <a href=\"https:\/\/prowad.pro\/fr\/ai-integration-education-implementation-framework\/\">implementing AI in education<\/a>. Schools can also use the <a href=\"https:\/\/prowad.pro\/fr\/evaluate-ai-learning-tools-schools-evidence-checklist\/\">evidence checklist for evaluating AI learning tools<\/a> when deciding which platforms deserve a place in staff development.<\/p>\n<h2>How to measure whether training is working<\/h2>\n<p>Attendance, satisfaction and licence activation are useful implementation indicators, but they are not sufficient evidence of competence. A more balanced evaluation can track four layers:<\/p>\n<ul>\n<li><strong>Adoption:<\/strong> Are teachers using the approved tool for relevant professional tasks, and are they using it voluntarily rather than merely logging in?<\/li>\n<li><strong>Prompt quality:<\/strong> Can teachers provide clear objectives, learner context, constraints and curriculum information in a new task?<\/li>\n<li><strong>Evaluation quality:<\/strong> Can they identify inaccurate, biased, incomplete or pedagogically unsuitable outputs and explain why they should be changed?<\/li>\n<li><strong>Classroom value:<\/strong> Do reviewed lesson materials, student work and classroom observations show a meaningful improvement in the learning process? This requires separate evidence; it cannot be inferred from better prompts alone.<\/li>\n<\/ul>\n<p>Where possible, use anonymized samples, transparent rubrics and blind scoring by more than one reviewer. Compare a baseline task with a different follow-up task so that teachers demonstrate transfer rather than memorization. Record the time required as one outcome, but do not make time saving the only definition of success.<\/p>\n<h2>What school leaders should take from the evidence<\/h2>\n<p>The central lesson is simple: an AI licence is an input, not a professional-development strategy. Teachers need a supported opportunity to practise, receive feedback, inspect failures and apply their subject expertise to generated material. The Peru trial suggests that even a relatively compact intervention can improve some of these capabilities when it combines practical training with continued support.<\/p>\n<p>But leaders should resist turning the result into a promise that one day of training will transform teaching or raise student achievement. The trial does not establish either claim. It establishes a narrower and useful point: teachers can become better at prompting and evaluating generative AI, and those skills may develop even when measurable time savings do not appear.<\/p>\n<p>For related guidance on when AI should enter lesson design, see <a href=\"https:\/\/prowad.pro\/fr\/when-should-teachers-use-ai-delayed-use-study\/\">when teachers should use AI<\/a>. For the wider division of responsibilities between educators and AI systems, read <a href=\"https:\/\/prowad.pro\/fr\/teacher-ai-collaboration-education-models\/\">four models of teacher\u2013AI collaboration<\/a>.<\/p>\n<h2>Conclusion: train teachers to judge, not just to prompt<\/h2>\n<p>The most useful measure of teacher AI readiness is not how quickly an educator can generate a worksheet. It is whether the educator can explain what the material is meant to achieve, recognize when the AI has failed, improve the result and decide when not to use the tool.<\/p>\n<p>The World Bank trial offers evidence that these skills can be strengthened through a package of practice-oriented training and follow-up. It also offers a caution: increased adoption does not automatically save time, and stronger AI fluency does not yet prove better student learning. Schools should therefore invest in professional judgment, measure it directly and evaluate classroom outcomes separately.<\/p>\n<h2>References and further reading<\/h2>\n<ul>\n<li>Glave, C. Z., Lopez, C., Molina, E., &amp; Z\u00e1rate, R. A. (2026). <a href=\"https:\/\/documents.worldbank.org\/en\/publication\/documents-reports\/documentdetail\/099309210012625906\" target=\"_blank\" rel=\"noopener\">Not Faster, but Better: Generative AI and Teacher Skill in Peru<\/a>. World Bank Policy Research Working Paper 11476, 1 October 2026. DOI: 10.1596\/1813-9450-11476.<\/li>\n<li>World Bank. (2026). <a href=\"https:\/\/reproducibility.worldbank.org\/catalog\/608\" target=\"_blank\" rel=\"noopener\">Reproducibility package for Not Faster, but Better<\/a>.<\/li>\n<li>Shenoy, P., &amp; Saarela, M. (2026). <a href=\"https:\/\/link.springer.com\/article\/10.1007\/s44217-026-01579-7\" target=\"_blank\" rel=\"noopener\">A critical review and actionable framework for integrating generative AI into teacher professional development<\/a>. <em>Discover Education<\/em>, 5, 570. https:\/\/doi.org\/10.1007\/s44217-026-01579-7.<\/li>\n<li>UNESCO. (2024). <a href=\"https:\/\/www.unesco.org\/en\/articles\/ai-competency-framework-teachers\" target=\"_blank\" rel=\"noopener\">AI Competency Framework for Teachers<\/a>.<\/li>\n<\/ul>\n<p><em>Editorial note: The main trial discussed here is a World Bank working paper, not a peer-reviewed journal article. The five-competency model and four-week pilot are practical recommendations for local testing, not interventions whose effects have been established by this trial.<\/em><\/p>","protected":false},"excerpt":{"rendered":"<p>A World Bank randomized trial with 1,269 teachers found AI training improved prompt quality and output judgment\u2014but did not save time or prove student learning gains.<\/p>","protected":false},"author":1,"featured_media":425,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[24],"tags":[],"class_list":["post-424","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-learning-training"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.3 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Teacher AI Evaluation Skills: 1,269-Teacher Trial<\/title>\n<meta name=\"description\" content=\"A World Bank trial with 1,269 teachers found AI training improved prompt quality and output judgment\u2014but did not save time or prove student learning gains.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/prowad.pro\/fr\/teacher-ai-evaluation-skills-world-bank-trial\/\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Teacher AI Evaluation Skills: 1,269-Teacher Trial\" \/>\n<meta property=\"og:description\" content=\"A World Bank trial with 1,269 teachers found AI training improved prompt quality and output judgment\u2014but did not save time or prove student learning gains.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/prowad.pro\/fr\/teacher-ai-evaluation-skills-world-bank-trial\/\" \/>\n<meta property=\"og:site_name\" content=\"prowad.pro\" \/>\n<meta property=\"article:published_time\" content=\"2026-10-31T18:04:35+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/prowad.pro\/wp-content\/uploads\/2026\/10\/Teacher-AI-Evaluation-Skills-Training.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1080\" \/>\n\t<meta property=\"og:image:height\" content=\"608\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"wadiagounine\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"wadiagounine\" \/>\n\t<meta name=\"twitter:label2\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data2\" content=\"11 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/\"},\"author\":{\"name\":\"wadiagounine\",\"@id\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/#\\\/schema\\\/person\\\/db804c968e78b362ba5e5eb5823d1fb1\"},\"headline\":\"Teacher AI Evaluation Skills: What a 1,269-Teacher Trial Found\",\"datePublished\":\"2026-10-31T18:04:35+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/\"},\"wordCount\":2311,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/prowad.pro\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/Teacher-AI-Evaluation-Skills-Training.jpg\",\"articleSection\":[\"Learning &amp; Training\"],\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/\",\"url\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/\",\"name\":\"Teacher AI Evaluation Skills: 1,269-Teacher Trial\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/prowad.pro\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/Teacher-AI-Evaluation-Skills-Training.jpg\",\"datePublished\":\"2026-10-31T18:04:35+00:00\",\"description\":\"A World Bank trial with 1,269 teachers found AI training improved prompt quality and output judgment\u2014but did not save time or prove student learning gains.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/#breadcrumb\"},\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/#primaryimage\",\"url\":\"https:\\\/\\\/prowad.pro\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/Teacher-AI-Evaluation-Skills-Training.jpg\",\"contentUrl\":\"https:\\\/\\\/prowad.pro\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/Teacher-AI-Evaluation-Skills-Training.jpg\",\"width\":1080,\"height\":608,\"caption\":\"Photo by Vitaly Gariev on Unsplash.\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/prowad.pro\\\/teacher-ai-evaluation-skills-world-bank-trial\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/prowad.pro\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Teacher AI Evaluation Skills: What a 1,269-Teacher Trial Found\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/#website\",\"url\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/\",\"name\":\"prowad.pro\",\"description\":\"Practical AI for people who teach, train and learn\",\"publisher\":{\"@id\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/#organization\",\"name\":\"prowad.pro\",\"url\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/prowad.pro\\\/wp-content\\\/uploads\\\/2026\\\/08\\\/cropped-cropped-387213115_210502682049592_4451028591485107083_n.png\",\"contentUrl\":\"https:\\\/\\\/prowad.pro\\\/wp-content\\\/uploads\\\/2026\\\/08\\\/cropped-cropped-387213115_210502682049592_4451028591485107083_n.png\",\"width\":2046,\"height\":681,\"caption\":\"prowad.pro\"},\"image\":{\"@id\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/#\\\/schema\\\/logo\\\/image\\\/\"}},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/prowadpro-cf4be0.ingress-erytho.ewp.live\\\/#\\\/schema\\\/person\\\/db804c968e78b362ba5e5eb5823d1fb1\",\"name\":\"wadiagounine\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/735c24828b869896fa7c15318b76e3677e85cc4a11ed5a9e159fd8f0c55c7c89?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/735c24828b869896fa7c15318b76e3677e85cc4a11ed5a9e159fd8f0c55c7c89?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/735c24828b869896fa7c15318b76e3677e85cc4a11ed5a9e159fd8f0c55c7c89?s=96&d=mm&r=g\",\"caption\":\"wadiagounine\"},\"sameAs\":[\"https:\\\/\\\/prowad.pro\"],\"url\":\"https:\\\/\\\/prowad.pro\\\/fr\\\/author\\\/wadiagounine\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Teacher AI Evaluation Skills: 1,269-Teacher Trial","description":"A World Bank trial with 1,269 teachers found AI training improved prompt quality and output judgment\u2014but did not save time or prove student learning gains.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/prowad.pro\/fr\/teacher-ai-evaluation-skills-world-bank-trial\/","og_locale":"fr_FR","og_type":"article","og_title":"Teacher AI Evaluation Skills: 1,269-Teacher Trial","og_description":"A World Bank trial with 1,269 teachers found AI training improved prompt quality and output judgment\u2014but did not save time or prove student learning gains.","og_url":"https:\/\/prowad.pro\/fr\/teacher-ai-evaluation-skills-world-bank-trial\/","og_site_name":"prowad.pro","article_published_time":"2026-10-31T18:04:35+00:00","og_image":[{"width":1080,"height":608,"url":"https:\/\/prowad.pro\/wp-content\/uploads\/2026\/10\/Teacher-AI-Evaluation-Skills-Training.jpg","type":"image\/jpeg"}],"author":"wadiagounine","twitter_card":"summary_large_image","twitter_misc":{"\u00c9crit par":"wadiagounine","Dur\u00e9e de lecture estim\u00e9e":"11 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/#article","isPartOf":{"@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/"},"author":{"name":"wadiagounine","@id":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/#\/schema\/person\/db804c968e78b362ba5e5eb5823d1fb1"},"headline":"Teacher AI Evaluation Skills: What a 1,269-Teacher Trial Found","datePublished":"2026-10-31T18:04:35+00:00","mainEntityOfPage":{"@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/"},"wordCount":2311,"commentCount":0,"publisher":{"@id":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/#organization"},"image":{"@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/#primaryimage"},"thumbnailUrl":"https:\/\/prowad.pro\/wp-content\/uploads\/2026\/10\/Teacher-AI-Evaluation-Skills-Training.jpg","articleSection":["Learning &amp; Training"],"inLanguage":"fr-FR","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/","url":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/","name":"Teacher AI Evaluation Skills: 1,269-Teacher Trial","isPartOf":{"@id":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/#website"},"primaryImageOfPage":{"@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/#primaryimage"},"image":{"@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/#primaryimage"},"thumbnailUrl":"https:\/\/prowad.pro\/wp-content\/uploads\/2026\/10\/Teacher-AI-Evaluation-Skills-Training.jpg","datePublished":"2026-10-31T18:04:35+00:00","description":"A World Bank trial with 1,269 teachers found AI training improved prompt quality and output judgment\u2014but did not save time or prove student learning gains.","breadcrumb":{"@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/#breadcrumb"},"inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/"]}]},{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/#primaryimage","url":"https:\/\/prowad.pro\/wp-content\/uploads\/2026\/10\/Teacher-AI-Evaluation-Skills-Training.jpg","contentUrl":"https:\/\/prowad.pro\/wp-content\/uploads\/2026\/10\/Teacher-AI-Evaluation-Skills-Training.jpg","width":1080,"height":608,"caption":"Photo by Vitaly Gariev on Unsplash."},{"@type":"BreadcrumbList","@id":"https:\/\/prowad.pro\/teacher-ai-evaluation-skills-world-bank-trial\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/prowad.pro\/"},{"@type":"ListItem","position":2,"name":"Teacher AI Evaluation Skills: What a 1,269-Teacher Trial Found"}]},{"@type":"WebSite","@id":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/#website","url":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/","name":"prowad.pro","description":"Practical AI for people who teach, train and learn","publisher":{"@id":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"fr-FR"},{"@type":"Organization","@id":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/#organization","name":"prowad.pro","url":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/","logo":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/#\/schema\/logo\/image\/","url":"https:\/\/prowad.pro\/wp-content\/uploads\/2026\/08\/cropped-cropped-387213115_210502682049592_4451028591485107083_n.png","contentUrl":"https:\/\/prowad.pro\/wp-content\/uploads\/2026\/08\/cropped-cropped-387213115_210502682049592_4451028591485107083_n.png","width":2046,"height":681,"caption":"prowad.pro"},"image":{"@id":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/#\/schema\/logo\/image\/"}},{"@type":"Person","@id":"https:\/\/prowadpro-cf4be0.ingress-erytho.ewp.live\/#\/schema\/person\/db804c968e78b362ba5e5eb5823d1fb1","name":"wadiagounine","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/secure.gravatar.com\/avatar\/735c24828b869896fa7c15318b76e3677e85cc4a11ed5a9e159fd8f0c55c7c89?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/735c24828b869896fa7c15318b76e3677e85cc4a11ed5a9e159fd8f0c55c7c89?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/735c24828b869896fa7c15318b76e3677e85cc4a11ed5a9e159fd8f0c55c7c89?s=96&d=mm&r=g","caption":"wadiagounine"},"sameAs":["https:\/\/prowad.pro"],"url":"https:\/\/prowad.pro\/fr\/author\/wadiagounine\/"}]}},"brizy_media":[],"_links":{"self":[{"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/posts\/424","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/comments?post=424"}],"version-history":[{"count":1,"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/posts\/424\/revisions"}],"predecessor-version":[{"id":427,"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/posts\/424\/revisions\/427"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/media\/425"}],"wp:attachment":[{"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/media?parent=424"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/categories?post=424"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prowad.pro\/fr\/wp-json\/wp\/v2\/tags?post=424"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}