[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"blog-article-navychky-spivpratsi-liudyny-z-ai":3},{"uuid":4,"title":5,"subtitle":6,"content":7,"preview":386,"date_published":391,"date_pinned":391,"img":392,"alias":393,"is_published":394,"is_pinned":395,"type":13,"category":396,"category_name":397,"category_alias":398,"background_color":399,"preview_icon":13,"category_value":400,"external_link":13},"385df278-263d-4986-a236-ce2f479b0754","Human–AI Collaboration Skills: What to Include in Curricula","Which skills are needed to work with AI, and how to turn them into practical assignments, outcomes, and curriculum criteria.",{"type":8,"content":9},"doc",[10,18,23,31,43,48,54,59,64,70,75,80,86,91,96,102,107,112,118,123,128,134,139,144,150,161,166,172,177,182,188,193,206,212,217,222,228,233,238,244,249,254,260,265,270,275,281,286,291,297,302,307,313,326,331,337],{"type":11,"attrs":12,"content":14},"paragraph",{"align":13},null,[15],{"text":16,"type":17},"A student can get a convincing answer from AI but cannot explain which data were used, what was checked, or why the result suits the task. Formally, they have mastered the tool, yet their collaboration remains unreliable. In a professional setting, they must also define the objective, choose an acceptable form of assistance, notice errors, and hand the result to another person. These actions should become visible learning outcomes, instead of limiting a program to a selection of effective prompts for one service.","text",{"type":11,"attrs":19,"content":20},{"align":13},[21],{"text":22,"type":17},"Collaboration with AI combines subject knowledge, an understanding of the tool's limitations, and work organization. It does not replace the course's core competence. A future analyst must understand data, an editor must understand meaning and evidence, and an instructor must understand the learning objective. AI changes how particular actions are performed, but responsibility for the result's suitability remains part of human work. A program should practice this connection through specific situations with a verifiable outcome.",{"type":24,"attrs":25,"content":28},"heading",{"align":13,"level":26,"id":27},2,"heading_2_1",[29],{"text":30,"type":17},"Turn Broad Literacy into Observable Actions",{"type":11,"attrs":32,"content":33},{"align":13},[34,41],{"text":35,"type":17,"marks":36},"UNESCO's AI competency framework for students",[37],{"type":38,"attrs":39},"link",{"href":40,"title":13},"https:\u002F\u002Fwww.unesco.org\u002Fen\u002Farticles\u002Fai-competency-framework-students",{"text":42,"type":17}," covers a human-centered mindset, ethics, AI techniques and applications, and system design. It provides structural guidance, rather than a ready-made plan for every professional course. The team must select relevant outcomes for its audience. For example, participants should be able to identify information missing from a task and stop working until they obtain it, rather than merely define a term in a test.",{"type":11,"attrs":44,"content":45},{"align":13},[46],{"text":47,"type":17},"For each competency, specify an action, its conditions, and evidence. ‘Think critically’ is too broad to assess in one lesson. ‘Identify an unsupported claim in an answer and find a source for evaluating it’ is already observable. Include a situation in which the right choice is not to obtain an answer: the data are insufficient, the constraints conflict, or a decision falls outside the student's authority. The ability to reject an appealing result should have the same educational value as successfully using assistance.",{"type":24,"attrs":49,"content":51},{"align":13,"level":26,"id":50},"heading_2_2",[52],{"text":53,"type":17},"Teach Students to Define the Task Before Opening the Chat",{"type":11,"attrs":55,"content":56},{"align":13},[57],{"text":58,"type":17},"Offer a short work request with an unclear objective: ‘Prepare better material for beginners.’ Students must establish who the material is for, which action it supports, what the constraints are, and how quality will be checked. Only then should they turn to the tool. Otherwise, the model will fill gaps with its own assumptions, and participants will treat the finished text as a solved problem. The exercise practices task definition, a skill that transfers between services and remains necessary after an interface changes.",{"type":11,"attrs":60,"content":61},{"align":13},[62],{"text":63,"type":17},"Students submit a brief statement of the objective and acceptance criteria. The instructor checks whether those criteria distinguish a useful version from merely polished prose. If a criterion says ‘interesting and modern,’ ask for specifics. For example, will beginners understand the required action without further explanation? Is there an example of its use? Does the difficulty match their prior preparation? This starting point makes subsequent prompts meaningful and provides a basis for checking what AI proposes.",{"type":24,"attrs":65,"content":67},{"align":13,"level":26,"id":66},"heading_2_3",[68],{"text":69,"type":17},"Divide Work into Parts Requiring Different Levels of Oversight",{"type":11,"attrs":71,"content":72},{"align":13},[73],{"text":74,"type":17},"One task may include exploring options, calculation, interpretation, and a final decision. There is no need to delegate every part to the tool in the same way. Students must determine where AI should suggest a draft, where an ordinary calculator is appropriate, and where subject-specific judgment is needed. For example, they can obtain several ways to explain something while independently checking an example's correctness. Mapping the parts reveals which results are intermediate and should not automatically move to the next stage without review.",{"type":11,"attrs":76,"content":77},{"align":13},[78],{"text":79,"type":17},"Set a time-limited exercise and ask students to justify their allocation of work. Assess the suitability of their choices rather than the maximum number of AI features used. A student may decline to use the tool for a brief action if verification would take longer than doing it independently. Another may use it to compare alternatives but not calculate the final metric. Both decisions can be sound if they fit the objective, risk level, and actual ability to check the result under the given conditions.",{"type":24,"attrs":81,"content":83},{"align":13,"level":26,"id":82},"heading_2_4",[84],{"text":85,"type":17},"Build the Habit of Providing Sufficient Context",{"type":11,"attrs":87,"content":88},{"align":13},[89],{"text":90,"type":17},"A work prompt should contain the necessary facts, constraints, and expected format, without unnecessary personal data. Create a fictional situation with several important parameters for a learning exercise. Ask students to separate what helps solve the task from incidental details. This combines subject-specific thinking with responsible information handling. Participants should be able to explain why they included a particular field, rather than simply pasting the entire available document into the chat without selection.",{"type":11,"attrs":92,"content":93},{"align":13},[94],{"text":95,"type":17},"It is useful to compare two prompts that differ only in a missing critical parameter. Have students predict how this gap will affect the answer, then test the versions. The objective is to understand the relationship between a condition and a result. There is no need to memorize a magic phrase. Context quality is assessed by whether the tool receives a sufficient basis for the required work and whether places where information is still insufficient for a particular conclusion remain visible.",{"type":24,"attrs":97,"content":99},{"align":13,"level":26,"id":98},"heading_2_5",[100],{"text":101,"type":17},"Teach Independent Ways to Verify an Answer",{"type":11,"attrs":103,"content":104},{"align":13},[105],{"text":106,"type":17},"Asking the same model to agree with its own text is insufficient confirmation. Students must be able to choose another route: open a primary source, repeat a calculation, test an example against reference data, or consult a subject-specific rule. Define in advance what counts as evidence in the exercise. Participants then practice verification, rather than producing another convincing explanation for an answer already accepted that adds no new information about its correctness.",{"type":11,"attrs":108,"content":109},{"align":13},[110],{"text":111,"type":17},"Use a draft with several types of flaws: an incorrect fact, an omitted condition, and an inappropriate conclusion. These require different verification tasks. Finding a source may correct a fact but does not guarantee that the conclusion follows logically. Students must identify the type of problem and how to fix it. Assess substantiated changes and uncertainties that remain explicit. If an unresolved condition remains after editing, honestly identifying that limit is better than confident prose that quietly invents the missing answer.",{"type":24,"attrs":113,"content":115},{"align":13,"level":26,"id":114},"heading_2_6",[116],{"text":117,"type":17},"Develop Subject-Specific Editing Skills",{"type":11,"attrs":119,"content":120},{"align":13},[121],{"text":122,"type":17},"AI can propose text that sounds complete but does not serve the learning product's objective. Ask students to shorten an explanation without losing a critical condition, or adapt it for a different audience. Ask them to show what changed and why. Preserving meaning is essential here. Removing every qualification for readability sometimes distorts a rule, while adding numerous terms may merely conceal the absence of a clear causal connection between statements.",{"type":11,"attrs":124,"content":125},{"align":13},[126],{"text":127,"type":17},"Use two versions and a specific reader for verification. Have another student try to perform the action using the edited instructions and identify points of uncertainty. The author then obtains evidence of the text's suitability, rather than only a stylistic assessment. The tool must not become the final judge of its own result. AI can suggest a version, but fitness for the assignment is established through meaning, sources, and the user's actual ability to perform the expected action under the intended learning conditions.",{"type":24,"attrs":129,"content":131},{"align":13,"level":26,"id":130},"heading_2_7",[132],{"text":133,"type":17},"Include Work with Uncertainty",{"type":11,"attrs":135,"content":136},{"align":13},[137],{"text":138,"type":17},"Professional tasks cannot always be reduced to one exact number. Students must distinguish known facts, assumptions, forecasts, and proposals. For an exercise, provide incomplete materials and ask for possible solutions with conditions for their use. Prohibit adding missing facts without labeling them. This format reveals whether participants can work within limitations or use the fluency of AI-generated text to conceal the lack of a necessary basis for a definite conclusion.",{"type":11,"attrs":140,"content":141},{"align":13},[142],{"text":143,"type":17},"Assess the quality of the next question. Sometimes the most useful result is correctly identifying which information to obtain from the client or instructor. Ask students to choose one clarification from several possibilities and explain its effect on the decision. This helps distinguish necessary verification from endless clarification of minor details. Competent collaboration with AI includes knowing when there is sufficient basis to proceed and when to stop because further generation would merely conceal a lack of knowledge.",{"type":24,"attrs":145,"content":147},{"align":13,"level":26,"id":146},"heading_2_8",[148],{"text":149,"type":17},"Practice Responsible Data Handling",{"type":11,"attrs":151,"content":152},{"align":13},[153,159],{"text":154,"type":17,"marks":155},"UNESCO's guidance on generative AI",[156],{"type":38,"attrs":157},{"href":158,"title":13},"https:\u002F\u002Fwww.unesco.org\u002Fen\u002Farticles\u002Fguidance-generative-ai-education-and-research",{"text":160,"type":17}," places human responsibility and appropriate conditions of use at the center. A curriculum should turn this into a specific exercise in selecting materials. Provide a mixed set: a public description, fictional records, and a document treated as confidential for the exercise. Students determine an acceptable route for each piece and explain why the entire set should not automatically be uploaded to any available service.",{"type":11,"attrs":162,"content":163},{"align":13},[164],{"text":165,"type":17},"Do not limit the lesson to a general statement about privacy. Have participants prepare a minimal version of the task without unnecessary details, check the tool's terms, and identify the responsible person for a disputed case. It is important to distinguish the technical ability to attach a file from the right to use it that way. Fictional data suffice for assessment. Real students' stories are unnecessary to demonstrate the ability to identify boundaries, remove unnecessary fields, and choose an agreed workflow.",{"type":24,"attrs":167,"content":169},{"align":13,"level":26,"id":168},"heading_2_9",[170],{"text":171,"type":17},"Teach Students to Hand Results to a Team",{"type":11,"attrs":173,"content":174},{"align":13},[175],{"text":176,"type":17},"Collaboration does not end in a personal chat. Colleagues need a result they can understand and verify without reading the entire prompt history. Students must prepare a concise description of the objective, materials used, checks performed, and open questions. This is not a formal declaration about AI; it passes on the basis for the next action. If another participant cannot tell what has already been confirmed, they will either duplicate work or trust an unverified passage as an established fact.",{"type":11,"attrs":178,"content":179},{"align":13},[180],{"text":181,"type":17},"Organize a paired exercise: one participant prepares material with AI, and the other accepts it against agreed criteria. They then exchange roles. Assess how many hidden assumptions are discovered during handover and whether the source of an important claim can be traced. Do not require every technical message to be retained unnecessarily. The necessary degree of transparency depends on the ability to verify the content and continue the work, rather than the number of appendix pages or screenshots of the interface used.",{"type":24,"attrs":183,"content":185},{"align":13,"level":26,"id":184},"heading_2_10",[186],{"text":187,"type":17},"Check the Ability to Work Without Assistance",{"type":11,"attrs":189,"content":190},{"align":13},[191],{"text":192,"type":17},"Part of the course should remain an independent demonstration of foundational knowledge. If students do not understand the subject's logic, they will struggle to notice plausible errors made by the tool. This does not mean banning AI at every stage. State honestly what each assignment assesses: independent mastery of the basics or the ability to organize assisted work. Mixing these outcomes makes feedback more difficult and creates an impression of progress unsupported by independent verification.",{"type":11,"attrs":194,"content":195},{"align":13},[196,198,204],{"text":197,"type":17},"After collaborative work, set a short new task without the same ready-made example. Ask students to explain a key decision or correct a deliberate error. ",{"text":199,"type":17,"marks":200},"CMU's principles of learning",[201],{"type":38,"attrs":202},{"href":203,"title":13},"https:\u002F\u002Fwww.cmu.edu\u002Fteaching\u002Fprinciples\u002Flearning.html",{"text":205,"type":17}," emphasize component skills, their integration, and the conditions for their use. For an AI-focused program, this means assessing not only an impressive collaborative product but also what participants understood, can reproduce, and can apply in another context after finishing work with the assistant.",{"type":24,"attrs":207,"content":209},{"align":13,"level":26,"id":208},"heading_2_11",[210],{"text":211,"type":17},"Build a Sequence with Several Levels",{"type":11,"attrs":213,"content":214},{"align":13},[215],{"text":216,"type":17},"At the introductory level, students work with given conditions and a short answer whose facts are easy to check. They then clarify the task themselves, choose a form of assistance, and compare alternatives. At a more advanced level, they organize a team process involving several sources and handover of the result. Every level introduces new responsibility. Increasing prompt length or the number of services does not itself demonstrate competency development if participants are not making more substantive and better-founded decisions.",{"type":11,"attrs":218,"content":219},{"align":13},[220],{"text":221,"type":17},"An end-to-end assignment might involve preparing an explanation for a beginner based on an open source. Students first identify the audience, then obtain alternative structures, check claims, edit an example, and test clarity with a classmate. Finally, they explain what they did themselves and which assistance they used. This is a project scenario rather than ready-made evidence of the method's effectiveness. The course team must verify whether the sequence supports its subject objective and the circumstances of the particular audience.",{"type":24,"attrs":223,"content":225},{"align":13,"level":26,"id":224},"heading_2_12",[226],{"text":227,"type":17},"Assess the Process Against a Few Meaningful Criteria",{"type":11,"attrs":229,"content":230},{"align":13},[231],{"text":232,"type":17},"A rubric can cover task clarity, the reasoning behind work allocation, verification quality, and the suitability of the result handed over. For each criterion, describe observable differences between levels. For example, weak verification repeats a prompt to the model; strong verification uses an independent basis and records the conclusion's limits. Do not try to assess an invisible thinking style from the text's confidence. You need materials and actions that students actually demonstrated and can explain in a brief conversation with the instructor.",{"type":11,"attrs":234,"content":235},{"align":13},[236],{"text":237,"type":17},"Include critical errors that style cannot compensate for: an invented source, a hidden assumption in a decisive calculation, or use of unapproved data. At the same time, do not penalize a justified decision not to use the tool. If an independent approach is better for the task, that may demonstrate sound judgment. The rubric should support responsible decisions rather than push students to demonstrate activity in a service merely to show AI use in every part of a learning assignment.",{"type":24,"attrs":239,"content":241},{"align":13,"level":26,"id":240},"heading_2_13",[242],{"text":243,"type":17},"Align Rules Across Instructors",{"type":11,"attrs":245,"content":246},{"align":13},[247],{"text":248,"type":17},"Students should understand where assistance is permitted, which involvement they must explain, and what remains independent work. Different subjects may have different rules, but their wording should not conflict without explanation. Prepare several typical examples: generating options, checking language, obtaining a complete solution, and analyzing a source. Ask instructors to assess them against shared principles. Discuss disagreements before they cause inconsistent grades or misunderstandings during the final assignment.",{"type":11,"attrs":250,"content":251},{"align":13},[252],{"text":253,"type":17},"The WeStudy LMS platform can help organize materials and the sequence of practice, while AI-use rules must be explicitly included in individual assignments. Do not rely on one general page students read at the start of the program. Alongside the work, state its objective, permitted assistance, and expected evidence. Participants then see the connection between the rule and the skill, and instructors have a basis for consistent assessment without having to guess exactly how submitted text was produced.",{"type":24,"attrs":255,"content":257},{"align":13,"level":26,"id":256},"heading_2_14",[258],{"text":259,"type":17},"Lesson Example: Checking an Explanation for a Beginner",{"type":11,"attrs":261,"content":262},{"align":13},[263],{"text":264,"type":17},"Take a short open-access resource on a familiar subject. Before the lesson, the instructor identifies three substantive conditions that the explanation must retain. Students receive the source and an audience profile, but no ready-made prompt. They first create their own plan and identify how AI could help. This initial version later makes it possible to compare the tool's contribution with their own intention and see which changes were accepted deliberately.",{"type":11,"attrs":266,"content":267},{"align":13},[268],{"text":269,"type":17},"Participants then ask for an alternative explanation and one example. They check the example against the source, rather than the answer's confidence. If the source cannot confirm a detail, they must remove it or label it unverified for further work. The instructor assesses whether all three decisive conditions remain. This matters more than the number of revisions: many successive prompts may only move the text further from its original meaning, even as each new version sounds smoother.",{"type":11,"attrs":271,"content":272},{"align":13},[273],{"text":274,"type":17},"Another student reads the result and answers a new brief question about applying the rule. If they make a mistake because a condition was omitted, the author corrects the material and explains the reason for the change. Finally, the author solves a similar task independently. The lesson thus yields several different forms of evidence: product quality, the ability to verify, and understanding of the subject's foundations. Analyze them separately so a successful text does not conceal the participant's lack of independent competence.",{"type":24,"attrs":276,"content":278},{"align":13,"level":26,"id":277},"heading_2_15",[279],{"text":280,"type":17},"Set a Team Task with Different Roles",{"type":11,"attrs":282,"content":283},{"align":13},[284],{"text":285,"type":17},"For a more advanced stage, divide preparation, verification, and acceptance of the result among three people. The first participant defines the task and prepares a draft with AI; the second checks the decisive claims; the third assesses suitability for the client. They must agree on which information moves between stages. Afterward, compare expectations with the actual handover: did the checker receive sources, did the editor see open questions, and were limitations lost when the final text was shortened?",{"type":11,"attrs":287,"content":288},{"align":13},[289],{"text":290,"type":17},"Change roles in the next exercise. Students who have accepted someone else's work can better see what their own explanations lack. Assessment should account for each person's contribution, rather than only the group's overall product. Offer a brief individual analysis of one disputed decision. This reveals whether participants understand the reasoning or simply agreed with a confident colleague or the model's wording. Team speed must not conceal unequal access to meaningful learning practice.",{"type":24,"attrs":292,"content":294},{"align":13,"level":26,"id":293},"heading_2_16",[295],{"text":296,"type":17},"Leave Room for Reasoned Disagreement",{"type":11,"attrs":298,"content":299},{"align":13},[300],{"text":301,"type":17},"Offer two different results, each with its own advantages and limitations. Students must choose one or produce a third, referring to the objective and evidence. Do not reveal beforehand which version was created by a person and which by AI. This exercise helps assess substance without automatically favoring a particular authorship. After revealing their origins, discuss whether the assessment changed and whether any new substantive basis emerged beyond the label identifying the answer's source.",{"type":11,"attrs":303,"content":304},{"align":13},[305],{"text":306,"type":17},"Instructors must also be ready to accept another justified decision. If the criteria allow several options, assess compliance with the conditions and the quality of the explanation. Do not turn collaboration with AI into guessing the desired wording. Participants must learn to defend a choice, acknowledge limitations, and change their position when a stronger basis emerges. This makes the course useful in situations with no ready-made model answer, where a responsible result must be agreed with other people.",{"type":24,"attrs":308,"content":310},{"align":13,"level":26,"id":309},"heading_2_17",[311],{"text":312,"type":17},"Update Tools While Retaining Learning Outcomes",{"type":11,"attrs":314,"content":315},{"align":13},[316,318,324],{"text":317,"type":17},"Services and feature names change faster than the fundamental actions of collaboration. ",{"text":319,"type":17,"marks":320},"The JRC maintains different editions of DigComp",[321],{"type":38,"attrs":322},{"href":323,"title":13},"https:\u002F\u002Fjoint-research-centre.ec.europa.eu\u002Fscientific-activities\u002Fkey-competences-lifelong-learning\u002Fdigital-competence-framework-digcomp\u002Fall-editions-digcomp-and-related-jrc-publications_en",{"text":325,"type":17},", helping people work with frameworks while taking the version into account. Similarly, identify the date and conditions of tool-specific instructions in your own program. Separately retain the competencies that should transfer: task definition, working with evidence, editing, responsibility, and handing a result to another person in an understandable form.",{"type":11,"attrs":327,"content":328},{"align":13},[329],{"text":330,"type":17},"Once per learning cycle, review exercises using the service's new version and check whether they still practice the required action. Automation may have removed a step students previously performed or introduced a new kind of error. If the WeStudy LMS platform is used for the program, update instructions and examples together with the criteria. Learning content should describe a person's ability to manage work with AI and assess its result in a subject context, rather than merely reproduce a sequence of clicks in a familiar product.",{"type":24,"attrs":332,"content":334},{"align":13,"level":26,"id":333},"heading_2_18",[335],{"text":336,"type":17},"Sources",{"type":338,"content":339},"bullet_list",[340,353,364,375],{"type":341,"content":342},"list_item",[343],{"type":11,"attrs":344,"content":345},{"align":13},[346,351],{"text":347,"type":17,"marks":348},"UNESCO: AI competency framework for students",[349],{"type":38,"attrs":350},{"href":40,"title":13},{"text":352,"type":17},".",{"type":341,"content":354},[355],{"type":11,"attrs":356,"content":357},{"align":13},[358,363],{"text":359,"type":17,"marks":360},"UNESCO: Guidance for generative AI in education and research",[361],{"type":38,"attrs":362},{"href":158,"title":13},{"text":352,"type":17},{"type":341,"content":365},[366],{"type":11,"attrs":367,"content":368},{"align":13},[369,374],{"text":370,"type":17,"marks":371},"JRC: All editions of DigComp and related publications",[372],{"type":38,"attrs":373},{"href":323,"title":13},{"text":352,"type":17},{"type":341,"content":376},[377],{"type":11,"attrs":378,"content":379},{"align":13},[380,385],{"text":381,"type":17,"marks":382},"Carnegie Mellon University: Learning Principles",[383],{"type":38,"attrs":384},{"href":203,"title":13},{"text":352,"type":17},{"type":8,"content":387},[388],{"type":11,"content":389},[390],{"text":6,"type":17},1791639531,"https:\u002F\u002Fcloud.cs3.westudy.ua\u002Ffile?path=westudy_online%2Fpublic%2Fwestudy%2F3f321f23-2649-4e64-b8bf-de0bc7643c1c.webp","navychky-spivpratsi-liudyny-z-ai",true,false,"1a2266a8-143b-428a-8320-d3f66d0468bb","Майбутнє освіти та навички","future-of-education-and-skills","light","news.articles"]