Competency catalogue

• Can explain what makes a system an AI system, and how it differs from conventional, explicitly programmed software.
• Can tell narrow AI from general AI, and name the main families of AI (rule-based, machine learning, generative).
• Can give examples of everyday AI and say what it can and cannot do.

An introduction whose aim is to explain what AI is: definitions, types, a short history, examples, and the contrast with traditional software, usually with a check of understanding.

A course or lesson that names AI in its title or opening line and moves straight on to a specific technique or tool.

• D08 Generative AI Literacy: D08 is about generative models specifically; D01 is about AI as a whole.
• D02 Machine Learning Fundamentals: D02 explains how machine learning learns; D01 stops at what AI is and what kinds exist.

• Can explain supervised, unsupervised and reinforcement learning and give an example of each.
• Can describe the steps of a machine learning project: data, features, training, evaluation, prediction.
• Can explain overfitting, and why a model is tested on data it did not see in training.

A lesson that explains learning paradigms or core techniques (regression, classification, clustering) to someone learning them, with examples and usually a worked notebook.

A text that calls a model 'supervised' or 'trained' in passing, or lists machine learning as a prerequisite.

• T06 AI Algorithm Design: T06 designs and implements algorithms for a task; D02 is understanding the paradigms.
• T05 Reinforcement Learning: T05 applies reinforcement learning in depth; D02 only places it among the paradigms.
• D01 AI Concept Understanding: D01 is what AI is in general; D02 is how machines learn from data.

• Can collect, clean and label data for training, and document where it came from.
• Can assess a dataset's quality: missing values, errors, balance and representativeness.
• Can explain how flaws in training data become flaws in a model.

Hands-on work on preparing data for AI — cleaning, labelling, exploring and checking quality — with the reasons each step matters for the model.

A notebook that loads a ready-made dataset and goes straight to modelling, or a text that names 'training data' without examining it.

• T09 Data Engineering for AI: T09 builds data pipelines and infrastructure; D03 is judging and preparing the data itself.
• T21 Bias Detection & Mitigation: T21 detects and mitigates bias in models; D03 checks data quality, bias in the data included.
• D16 Data Privacy in AI Contexts: D16 protects personal data; D03 is about its quality and fitness for training.

• Can write clear, structured prompts with a role, context, constraints and an output format.
• Can use techniques such as few-shot examples, step-by-step reasoning and iterative refinement.
• Can compare prompts and revise them according to the results.

A tutorial or lab on writing prompts: techniques explained, examples of better and worse prompts, and exercises that have the learner write and refine their own.

Code that sends a fixed prompt to an API without discussing it, or a text that says 'prompt the model'.

• D05 AI Output Evaluation: D05 judges what comes back; D04 shapes what goes in.
• D26 AI-Assisted Coding: D26 is working with coding assistants; D04 is prompting in general.
• D08 Generative AI Literacy: D08 is understanding what generative models can do; D04 is getting them to do it.

• Can check AI-generated content for factual accuracy against reliable sources.
• Can recognise hallucination, bias and inconsistency in a model's output.
• Can decide whether to accept, revise or reject an output, and justify the decision.

Material that teaches users to scrutinise generated output — accuracy, bias, reliability — with examples of failures and practice in checking them.

A warning such as 'AI can make mistakes' at the end of a tutorial, or a disclaimer that outputs should be verified.

• T10 AI Model Evaluation: T10 applies formal evaluation metrics to models; D05 is a user's critical reading of outputs.
• D15 AI & Information Verification: D15 verifies information in general, AI-made or not; D05 assesses what an AI produced.
• T21 Bias Detection & Mitigation: T21 measures and mitigates bias in models; D05 spots it in outputs.

• Can break a problem into steps that a machine, or an AI assistant, can carry out.
• Can abstract the essential features of a problem and express a procedure for it.
• Can decide which steps of a solution suit AI and which do not.

Teaching material on decomposition, abstraction and algorithms, with problems the learner works through step by step.

A tutorial that lists steps to follow without asking the learner to design them.

• T25 Advanced AI Problem Formulation: T25 formalises complex problems into AI models; D06 is the general habit of decomposing problems.
• T06 AI Algorithm Design: T06 designs machine learning algorithms; D06 is foundational algorithmic thinking.

• Can combine an AI system's recommendation with their own judgement and context.
• Can say when to follow, question or override an AI recommendation.
• Can explain the limits of decision support systems and the role of the human in them.

Material on using AI to inform decisions — decision support, recommendations, cases where human judgement must prevail — with scenarios the learner reasons through.

A recommender system built as a coding exercise, with nothing about the decisions people take with it.

• R18 AI-Supported Decision Agility: R18 is making rapid decisions with analytics; D07 is weighing AI insight against human judgement.
• D22 Responsible AI Use Mindset: D22 keeps a human in the loop for ethical reasons; D07 does it for the quality of the decision.
• R27 Human Agency & Oversight Strategy: R27 designs oversight for an organisation; D07 is the individual decision.

• Can explain what generative models produce and, broadly, how.
• Can describe their limitations: hallucination, bias, knowledge cut-off, cost.
• Can choose sensible uses for generative AI and recognise poor ones.

An introduction to generative AI for users: what large language and image models do, what they cannot do, and examples that show both.

A lab that calls a generative model through an API to build an application, with nothing on its capabilities or limits.

• T12 Generative AI Architecture: T12 explains the architecture inside generative models; D08 is what they can and cannot do.
• D01 AI Concept Understanding: D01 is AI in general; D08 is generative models.
• D04 AI Interaction & Prompting: D04 is prompting them well; D08 is understanding them.

• Can compare AI tools for a task on capability, cost, data handling and fit.
• Can decide whether a task needs AI at all.
• Can justify the choice of a tool to others.

Material that teaches how to choose among AI tools or services, with criteria, comparisons and cases where the learner makes the choice.

A tutorial that uses one tool without discussing why it was chosen.

• G03 Sustainable AI Model Selection: G03 chooses models on environmental criteria; D09 on fit for the task.
• G21 Sustainable AI Procurement: G21 evaluates vendors on sustainability; D09 selects tools for a task.
• D23 AI Workflow Integration: D23 embeds a chosen tool into a workflow; D09 is choosing it.

• Can divide a task between people and AI according to what each does well.
• Can design a workflow in which AI augments rather than replaces human work.
• Can recognise when collaboration with AI degrades the quality of the work.

Material on human–AI teaming: strengths of each, task allocation, examples of workflows that combine them, with exercises in planning such a division.

A statement that 'AI will assist humans' with no account of how the work is divided.

• D27 Working with AI Agents: D27 supervises autonomous agents; D10 divides work between people and AI.
• R11 Collaborative AI Problem Solving: R11 is cross-functional teams solving problems with AI; D10 is the human–AI pairing itself.

• Can produce text, images or media with AI and edit the result into their own work.
• Can iterate with an AI tool to improve a piece of content.
• Can acknowledge AI's contribution and respect rights in co-created content.

Practical material on creating content with AI — drafting, illustrating, editing — where the learner produces and revises something.

A demo that generates an image or a paragraph to show what a model can do.

• D25 AI & Creative Problem Solving: D25 uses AI to solve problems creatively; D11 produces content.
• D21 AI-Enhanced Communication: D21 adapts communication across languages and formats; D11 creates new content.

• Can explain why the opacity of some AI systems is a problem, and for whom.
• Can assess how transparent and explainable a given AI system is.
• Can ask the right questions about how an AI decision was reached.

Material that explains transparency and explainability as properties to evaluate — black boxes, explanations, documentation such as model cards — with cases to assess.

A passing reference to 'black box models' in a text about something else.

• T18 Explainable AI (XAI): T18 implements explanation techniques such as SHAP and LIME; D12 evaluates transparency.
• R27 Human Agency & Oversight Strategy: R27 designs human oversight; D12 assesses whether a system can be understood.
• T22 Responsible AI Development: T22 builds transparency into development; D12 judges it from outside.

• Can explain how computers process and generate human language, in broad terms.
• Can describe common NLP tasks: classification, sentiment, translation, speech, generation.
• Can recognise NLP in everyday applications and its typical failures.

An introductory lesson on how language technology works — tokens, models, tasks — aimed at understanding rather than building.

An NLP library imported in a notebook about another subject, or a list of NLP tasks in an overview.

• T03 NLP Engineering: T03 builds and fine-tunes NLP systems; D13 is understanding them.
• D21 AI-Enhanced Communication: D21 uses AI to communicate across languages; D13 explains how language processing works.

• Can explain how AI recognises and generates images, in broad terms.
• Can name computer vision applications and their limitations.
• Can recognise risks such as misclassification and synthetic images.

An introduction to computer vision for understanding — how images are represented, classified and generated — with examples and simple activities.

An image used in a notebook about another subject, or a vision library listed among dependencies.

• T04 Computer Vision Development: T04 builds and deploys vision systems; D14 is recognising and understanding them.
• D08 Generative AI Literacy: D08 covers image generation as one generative capability; D14 covers vision as a whole.

• Can check the credibility of sources and cross-reference claims.
• Can recognise synthetic media and other signs of manipulated content.
• Can use AI tools, critically, to support verification.

Material on detecting misinformation and disinformation — source checks, deepfakes, provenance — with practice on real or realistic examples.

A note that 'deepfakes exist' in a text about image generation.

• D05 AI Output Evaluation: D05 evaluates AI outputs; D15 verifies information wherever it comes from.
• D17 AI Safety & Security: D17 is security risks to AI systems; D15 is protecting people from false information.

• Can identify personal data in an AI workflow and the risks of processing it.
• Can apply principles such as minimisation, purpose limitation and consent.
• Can configure tools and practices that protect privacy when using AI.

Material on protecting personal data when using AI — GDPR principles, anonymisation, consent, privacy settings — with practice applying them to cases or code.

A footer link to a privacy statement, or a passing instruction not to paste secrets into a chatbot.

• G09 Eco-friendly Data Management: G09 minimises data for sustainability; D16 for privacy.
• D17 AI Safety & Security: D17 is security threats to AI; D16 is protection of personal data.
• D19 AI Regulatory Awareness: D19 is the AI Act's risk classes; D16 applies data protection principles.

• Can recognise attacks on AI systems: prompt injection, data poisoning, adversarial examples.
• Can explain the security risks of deploying AI and how they are mitigated.
• Can apply basic safeguards when building or using AI applications.

Material on AI security — threats, attacks, mitigations — with examples and, for builders, practice in defending an application.

A repository's standard security policy for reporting vulnerabilities, or a note to keep API keys secret.

• R25 Resilient AI Systems Design: R25 builds robust, fault-tolerant systems; D17 recognises security risks.
• D16 Data Privacy in AI Contexts: D16 protects personal data; D17 protects AI systems from attack.

• Can use AI to make digital content more accessible: captions, alt text, simplification.
• Can apply accessibility guidelines when designing with AI.
• Can recognise where AI tools exclude users with disabilities.

Material on accessibility with AI — assistive technology, accessible content, inclusive design — with practice producing or checking accessible material.

An alt attribute on an image, or a sentence promising that an app is accessible.

• D21 AI-Enhanced Communication: D21 adapts communication in general; D18 is access for people with disabilities.
• D11 Digital Content Co-creation: D11 creates content with AI; D18 makes it accessible.

• Can explain the EU AI Act's risk categories and give examples of each.
• Can identify the obligations that apply to a given AI system.
• Can recognise prohibited practices and high-risk uses.

Material that explains the AI Act — risk classes, obligations, prohibited practices — with cases the learner classifies.

A sentence noting that AI is increasingly regulated.

• R23 Regulatory Navigation for AI: R23 navigates regulation in a business; D19 understands the AI Act's classification.
• R27 Human Agency & Oversight Strategy: R27 designs oversight under Article 14; D19 is the regulation as a whole.
• D24 AI Impact Assessment: D24 assesses societal impact; D19 legal classification.

• Can use AI to translate, summarise or reformat a message for another audience.
• Can check AI-translated or adapted text for meaning and tone.
• Can choose formats and channels that AI can help produce.

Practical material on using AI to communicate — translation, plain language, transcription, adaptation to an audience — where the learner produces and checks such material.

A notice that a document was machine-translated, or a lesson on how translation models work.

• D13 NLP Awareness: D13 explains how language processing works; D21 uses AI to communicate.
• T03 NLP Engineering: T03 builds translation and NLP systems; D21 uses them.
• D18 AI & Accessibility: D18 is access for people with disabilities; D21 is communication across languages and formats.

• Can identify ethical risks in a concrete AI workflow.
• Can choose human-in-the-loop alternatives where AI should not decide alone.
• Can contest or override an AI output when it is wrong or unfair.

Material that works through responsible use of AI in real tasks — risks, safeguards, when to keep a human in the loop — through cases and reflection.

A closing paragraph urging readers to 'use AI responsibly', or a link to a responsible AI policy.

• T22 Responsible AI Development: T22 is responsible development by builders; D22 is responsible use.
• D07 AI-Augmented Decision Making: D07 keeps a human in the loop for decision quality; D22 for ethics.
• R15 AI Ethics in Entrepreneurship: R15 applies ethics to commercial deployment; D22 to everyday use.

• Can embed AI tools into a professional or educational workflow.
• Can redesign the steps of a workflow around what AI does well.
• Can evaluate whether the integration actually improves the work.

Material on integrating AI into everyday work processes, with examples of workflows before and after and practice designing one.

A CI workflow file, or a tool described as fitting 'into your workflow'.

• R14 Resilience Through Automation: R14 automates processes for resilience; D23 embeds AI into workflows.
• R26 AI-Powered Process Optimisation: R26 optimises organisational processes with AI; D23 integrates tools into one's own work.
• D09 AI Tools Selection: D09 chooses the tool; D23 builds it into the workflow.

• Can assess the societal, economic and personal effects of an AI system.
• Can weigh benefits against harms for different groups.
• Can use or design an impact assessment for an AI application.

Material on assessing AI's impact on society, the economy and individuals, with frameworks and cases the learner analyses.

An opening sentence on how AI is 'transforming society'.

• G04 AI Lifecycle Sustainability: G04 assesses environmental impact over a lifecycle; D24 societal and economic impact.
• R20 AI & Social Innovation: R20 applies AI for social good; D24 assesses impact.
• D19 AI Regulatory Awareness: D19 is legal classification; D24 is impact.

• Can use AI to generate ideas and explore alternatives to a problem.
• Can reframe a problem with AI's help and test new approaches.
• Can judge which AI-generated ideas are worth pursuing.

Material on creative problem solving with AI — ideation, reframing, prototyping — where the learner works a problem with AI as a partner.

A suggestion to 'brainstorm with ChatGPT' in a list of tips.

• D11 Digital Content Co-creation: D11 produces content; D25 solves problems.
• R10 AI & Innovation Management: R10 manages innovation in an organisation; D25 is individual creative problem solving.
• D06 Algorithmic Thinking: D06 decomposes problems systematically; D25 explores them creatively.

• Can use an AI assistant to draft, refactor and explain code.
• Can verify AI-generated code for correctness, security and licensing.
• Can decide when AI-assisted coding helps and when it hides problems.

Material on coding with AI assistants — prompting for code, reviewing and testing what comes back — with practice in a real coding task.

Code that a tutorial says was written with an assistant, or a list of editor extensions.

• T16 AI Programming Proficiency: T16 writes machine learning code in Python; D26 uses AI to write any code.
• T08 AI & Software Integration: T08 integrates AI into software; D26 uses AI while writing software.
• D04 AI Interaction & Prompting: D04 is prompting in general; D26 is coding with an assistant.

• Can configure an AI agent with tools, instructions and limits on its autonomy.
• Can supervise a multi-step agent run and inspect its trace.
• Can decide when an agent needs a human approval step.

Material on building or operating AI agents — tools, planning, multi-agent setups, supervision and traces — where the learner configures and runs an agent.

A passing reference to 'agentic AI' as a trend, or an agent framework listed among dependencies.

• T08 AI & Software Integration: T08 integrates tools and APIs into software; D27 configures and supervises agents.
• D10 Human-AI Collaboration: D10 divides work between people and AI; D27 bounds an agent's autonomy.
• T05 Reinforcement Learning: T05's agents learn from rewards; D27's agents use models and tools to do tasks.

• Can explain where AI's energy use and emissions come from: training, inference, data centres, hardware.
• Can compare the footprint of different AI activities in orders of magnitude.
• Can name the factors that raise or lower it: model size, hardware, location, usage.

Material that explains the environmental footprint of AI — energy, carbon, water — with figures, comparisons and their sources.

A sentence noting that training large models is expensive, in a text about something else.

• G06 Carbon Footprint Measurement: G06 measures and reports emissions; G01 understands where they come from.
• G04 AI Lifecycle Sustainability: G04 assesses the whole lifecycle; G01 the footprint of AI activity.

• Can apply energy-efficient practices when running AI workloads.
• Can choose hardware, settings and schedules that reduce energy use.
• Can explain the trade-off between performance and energy.

Practical material on running computation efficiently — efficient hardware use, batching, scheduling, avoiding waste — with exercises applying it.

An efficiency tip about a smaller batch size given for speed, with no word on energy.

• G05 Green Software for AI: G05 applies green software engineering to AI applications; G02 is energy-efficient computing in general.
• G15 AI-Driven Energy Optimization: G15 uses AI to manage an organisation's energy; G02 makes computing itself efficient.
• G22 Low-Carbon AI Deployment: G22 designs low-carbon deployment; G02 applies everyday efficiency.

• Can compare candidate models on their energy and resource cost as well as their accuracy.
• Can choose a smaller or more efficient model when it meets the need.
• Can justify a model choice on environmental grounds.

Material that teaches choosing models with their environmental cost in view — efficiency benchmarks, small models, trade-offs — with cases where the learner makes the choice.

Choosing among models for accuracy alone, or sklearn's model_selection module in code.

• G25 Green AI Innovation Practice: G25 proposes low-energy alternatives when designing; G03 chooses among existing models.
• G08 Model Compression & Optimization: G08 compresses a model; G03 selects an efficient one.
• G21 Sustainable AI Procurement: G21 evaluates vendors; G03 models.

• Can describe the environmental impact of each stage of an AI system's lifecycle.
• Can apply life-cycle assessment thinking to an AI project.
• Can identify the stage where impact is greatest for a given system.

Material on the lifecycle impact of AI — hardware, data, training, inference, disposal — using life-cycle assessment, with a case worked through.

A machine learning lifecycle diagram with no environmental content.

• G01 AI Environmental Footprint: G01 is the footprint of AI activity; G04 the whole lifecycle.
• G11 E-waste & Hardware Lifecycle: G11 is hardware and e-waste; G04 includes them among every stage.
• G06 Carbon Footprint Measurement: G06 measures emissions; G04 assesses impact across stages.

• Can apply green software principles to an AI application.
• Can make an application carbon-aware or more energy-efficient in its code and architecture.
• Can measure the effect of a change on energy use.

Material on green software engineering for AI applications — efficient code, carbon awareness, measuring software carbon intensity — with practice on an application.

A performance optimisation presented for speed or cost only.

• G02 Green Computing Principles: G02 is energy-efficient computing in general; G05 is software engineering for AI apps.
• G06 Carbon Footprint Measurement: G06 measures emissions; G05 reduces them through software design.
• G22 Low-Carbon AI Deployment: G22 is deployment strategy; G05 is the software itself.

• Can measure or estimate the energy use and emissions of an AI workload.
• Can use tools such as CodeCarbon or cloud carbon dashboards and read their output.
• Can report emissions with their method and assumptions.

Hands-on material on measuring AI emissions — tools, methods, emission factors — where the learner measures a workload and reports it.

A figure for a model's training emissions quoted from a paper.

• G01 AI Environmental Footprint: G01 understands the footprint; G06 measures it.
• G16 Sustainability Impact Reporting: G16 communicates impact in reports; G06 produces the measurements.
• G05 Green Software for AI: G05 reduces emissions through software; G06 measures them.

• Can explain how the energy mix of data centres affects AI's emissions.
• Can describe the role of renewable energy and its limits for AI infrastructure.
• Can compare regions or providers on the carbon intensity of their electricity.

Material on renewable energy for AI infrastructure — energy mix, power purchase agreements, location — with data the learner interprets.

A provider's claim to run on renewable energy, cited in passing.

• G18 Responsible Cloud Computing: G18 selects cloud services by sustainability; G07 understands renewable energy's role.
• G20 AI & Climate Action: G20 is AI's role in climate action; G07 the energy behind AI.

• Can apply pruning, quantisation or distillation to a model.
• Can measure the trade-off between a compressed model's size, speed and accuracy.
• Can choose a compression technique for a deployment constraint.

Hands-on material on model compression — pruning, quantisation, distillation — where the learner compresses a model and measures the result.

A quantised model downloaded and used as is, with nothing about the technique.

• G03 Sustainable AI Model Selection: G03 selects an efficient model; G08 makes a model efficient.
• G22 Low-Carbon AI Deployment: G22 designs low-carbon deployment; G08 reduces the model.
• T19 Edge AI & IoT Integration: T19 deploys to edge devices, often after compression; G08 is the compression.

• Can reduce the data an AI project stores and processes without losing what it needs.
• Can apply sustainable storage and processing practices: retention, deduplication, efficient formats.
• Can estimate the footprint of a data pipeline.

Material on sustainable data management for AI — minimisation, retention, efficient formats and processing — with practice applying it.

A data pipeline described for speed or cost only.

• D16 Data Privacy in AI Contexts: D16 minimises data for privacy; G09 for sustainability.
• T09 Data Engineering for AI: T09 builds data pipelines; G09 makes data management sustainable.

• Can use AI tools to monitor environmental conditions: air, water, land, climate.
• Can work with environmental data such as satellite imagery or sensor streams.
• Can interpret the results and their uncertainty.

Material that applies AI to environmental monitoring — remote sensing, sensors, earth observation — with data the learner analyses.

An example dataset about the weather used to teach a generic technique.

• G24 AI & Biodiversity: G24 is biodiversity specifically; G10 environmental monitoring in general.
• G20 AI & Climate Action: G20 is AI's role in climate action; G10 monitoring.

• Can explain the environmental impact of manufacturing and disposing of digital devices.
• Can describe embodied carbon and how it compares to operational emissions.
• Can name practices that extend hardware life and recover materials.

Material on e-waste and the hardware lifecycle — manufacturing, embodied carbon, repair, recycling — with figures and cases.

A note that GPUs are scarce, in a text about training.

• G04 AI Lifecycle Sustainability: G04 covers every lifecycle stage; G11 hardware and its disposal.
• G14 Circular Economy & AI: G14 uses AI for circular economy; G11 is the impact of devices.

• Can identify the environmental cost of everyday digital behaviour.
• Can adopt habits that reduce it: storage, streaming, devices, AI use.
• Can explain the reasoning behind those habits to others.

Material on sustainable digital habits — what costs energy in everyday use and what to change — with reflection on one's own practice.

A tip to close unused tabs, in a text about something else.

• G23 Digital Sobriety: G23 practises minimal, intentional use; G12 adopts lower-impact habits.
• G17 Green AI Advocacy: G17 promotes sustainable practices in communities; G12 is one's own behaviour.

• Can apply GreenComp competences such as systems thinking and futures literacy to choosing AI tools.
• Can weigh the sustainability consequences of using AI in a given context.
• Can frame a decision about AI as a sustainability problem.

Material that uses the GreenComp framework, or its competences, as a lens on AI use, with activities that apply it to real choices.

GreenComp cited in a reference list.

• G19 Environmental Ethics in AI: G19 integrates environmental ethics into AI ethics; G13 applies GreenComp competences.
• R12 EntreComp Lens for AI Opportunity: R12 applies EntreComp; G13 GreenComp.

• Can identify where AI supports circular economy principles: reuse, repair, recycling.
• Can describe AI applications such as material sorting or demand prediction for reuse.
• Can assess whether such an application reduces resource use overall.

Material on AI for the circular economy, with applications explained and a case the learner analyses or builds.

Circular economy named as an example sector in a list.

• G11 E-waste & Hardware Lifecycle: G11 is the impact of devices; G14 is AI supporting circularity.
• G15 AI-Driven Energy Optimization: G15 optimises energy; G14 materials.

• Can apply AI to forecast and optimise energy use in buildings or operations.
• Can work with energy consumption data and interpret a model's recommendations.
• Can evaluate the savings an optimisation achieves.

Material that applies AI to energy management — forecasting demand, optimising consumption — with data the learner models.

An energy dataset used to teach regression, with no energy management purpose.

• G02 Green Computing Principles: G02 makes computing efficient; G15 uses AI to manage energy.
• G07 Renewable Energy for AI: G07 is renewable energy for AI; G15 AI for energy.

• Can report the environmental impact of AI use clearly and with its method.
• Can choose metrics and frameworks for sustainability reporting.
• Can communicate uncertainty and avoid greenwashing.

Material on reporting AI's sustainability impact — metrics, frameworks, disclosure — with practice drafting a report.

A company's sustainability report cited as a source.

• G06 Carbon Footprint Measurement: G06 measures; G16 communicates the measurements.
• G17 Green AI Advocacy: G17 advocates; G16 reports.

• Can make the case for sustainable AI practices to a community or organisation.
• Can organise actions that raise awareness of AI's environmental cost.
• Can tailor arguments to different audiences.

Material on advocating sustainable AI — campaigns, community engagement, messaging — with activities that plan or carry out advocacy.

A call to 'think about the planet' at the end of a talk.

• G16 Sustainability Impact Reporting: G16 reports impact; G17 advocates change.
• G13 GreenComp Lens for AI: G13 applies a competence framework; G17 promotes practice.
• R24 Community of Practice Building: R24 builds communities of practice; G17 campaigns.

• Can compare cloud services and regions on sustainability criteria.
• Can configure workloads to run where and when electricity is cleaner.
• Can read and question providers' sustainability claims.

Material on choosing cloud services sustainably — regions, providers' data, carbon-aware scheduling — with a choice the learner makes and justifies.

Cloud deployment instructions that pick a region for latency or price.

• G21 Sustainable AI Procurement: G21 evaluates vendors in procurement; G18 selects cloud services.
• G07 Renewable Energy for AI: G07 explains renewable energy; G18 acts on it when choosing services.
• G22 Low-Carbon AI Deployment: G22 designs low-carbon deployment; G18 picks the cloud services.

• Can include environmental harms when assessing the ethics of an AI system.
• Can reason about fairness between generations and regions in AI's environmental cost.
• Can argue for environmental responsibility in AI decisions.

Material that treats environmental considerations as part of AI ethics — justice, responsibility, planetary limits — through cases and reflection.

An AI ethics overview that lists sustainability as one item among many.

• G13 GreenComp Lens for AI: G13 applies GreenComp competences; G19 is environmental ethics.
• T22 Responsible AI Development: T22 is responsible development in general; G19 adds the environment to it.
• D22 Responsible AI Use Mindset: D22 is responsible use in general; G19 is its environmental side.

• Can explain how AI contributes to climate mitigation and adaptation.
• Can name applications such as emissions forecasting, climate modelling and early warning.
• Can weigh AI's contribution against its own footprint.

Material on AI for climate action — applications, evidence, limits — with examples or data the learner analyses.

Climate change used as a motivating example in an unrelated lesson.

• G10 AI for Environmental Monitoring: G10 monitors the environment; G20 is AI's role in climate action.
• G15 AI-Driven Energy Optimization: G15 optimises energy in organisations; G20 addresses climate broadly.

• Can evaluate AI vendors with sustainability criteria.
• Can write sustainability requirements into a procurement.
• Can check vendors' environmental claims and evidence.

Material on sustainable procurement of AI — criteria, questions for vendors, evaluation — with a procurement case.

A vendor's sustainability page linked in passing.

• G18 Responsible Cloud Computing: G18 selects cloud services; G21 evaluates vendors in procurement.
• G03 Sustainable AI Model Selection: G03 selects models; G21 vendors.
• D09 AI Tools Selection: D09 selects tools for fit; G21 vendors for sustainability.

• Can design a deployment that minimises emissions: hardware, region, scaling, scheduling.
• Can estimate the emissions of serving a model at a given load.
• Can compare deployment options on carbon as well as cost.

Material on low-carbon deployment of AI — efficient serving, carbon-aware scheduling, right-sizing — with a deployment the learner designs.

Deployment instructions that scale a service for traffic, with nothing on emissions.

• T13 AI Systems Deployment: T13 deploys and monitors AI in production; G22 does it with minimal emissions.
• G08 Model Compression & Optimization: G08 compresses the model; G22 designs the deployment.
• G18 Responsible Cloud Computing: G18 chooses cloud services; G22 designs the whole deployment.

• Can decide when a digital or AI tool is not needed.
• Can practise intentional, minimal use of digital resources.
• Can explain digital sufficiency and its benefits.

Material on digital sobriety — sufficiency, minimal use, low-tech alternatives — with reflection on one's own use.

A passing remark about using fewer devices.

• G12 Sustainable Digital Practices: G12 adopts lower-impact habits; G23 uses less.
• G25 Green AI Innovation Practice: G25 designs frugal AI solutions; G23 is personal practice.

• Can describe AI applications for biodiversity: species identification, acoustic monitoring, habitat mapping.
• Can work with ecological data such as camera-trap images or audio.
• Can discuss the limits of AI in conservation.

Material that applies AI to biodiversity conservation, with data or a case the learner works on.

An animal image dataset used to teach image classification.

• G10 AI for Environmental Monitoring: G10 monitors the environment in general; G24 biodiversity.
• G20 AI & Climate Action: G20 is climate action; G24 conservation of species.

• Can propose a lower-energy or lower-data alternative when designing an AI solution.
• Can judge whether a simpler method meets the need.
• Can argue for frugal design to stakeholders.

Material on frugal AI design — simpler models, less data, non-AI alternatives — with design problems the learner works through.

A note that a smaller model was faster.

• G03 Sustainable AI Model Selection: G03 chooses among existing models; G25 proposes alternatives in design.
• G08 Model Compression & Optimization: G08 compresses a model; G25 rethinks the solution.
• D25 AI & Creative Problem Solving: D25 is creative problem solving in general; G25 aims at lower energy and data.

• Can use AI tools to scan for business or professional opportunities.
• Can turn trends found with AI into concrete opportunity statements.
• Can judge which AI-surfaced opportunities are worth pursuing.

Material on recognising opportunities with AI — scanning trends, spotting needs, testing ideas — with practice on a real or realistic market.

A line that AI 'creates new opportunities' in an introduction.

• R12 EntreComp Lens for AI Opportunity: R12 applies EntreComp to AI opportunities; R01 uses AI to find them.
• R13 AI Market Analysis: R13 analyses markets and competitors; R01 spots opportunities.
• R16 AI Value Proposition Design: R16 designs the value proposition; R01 finds the opportunity.

• Can analyse how AI changes their sector or profession.
• Can adjust a professional or business strategy in response.
• Can plan scenarios for different paces of AI adoption.

Material on adapting strategy to AI-driven change — analysis, scenarios, strategic options — with cases the learner works through.

A sentence that organisations need an AI strategy.

• R17 Adaptation to AI-Driven Change: R17 adapts one's own practice; R02 adjusts strategy.
• R05 AI Change Management: R05 leads organisational change; R02 sets the strategy.
• R08 Digital Transformation Leadership: R08 leads transformation; R02 adapts strategy.

• Can use AI and data to identify and prioritise risks.
• Can interpret risk predictions and their uncertainty.
• Can combine model output with expert judgement in a risk assessment.

Material that applies AI to risk identification — risk models, prediction, interpretation — with data or cases the learner analyses.

A risk register mentioned in a project template.

• D07 AI-Augmented Decision Making: D07 integrates AI into decisions generally; R03 identifies risks.
• R18 AI-Supported Decision Agility: R18 decides quickly; R03 assesses risk.
• R25 Resilient AI Systems Design: R25 builds resilient AI systems; R03 uses AI to find risks.

• Can combine entrepreneurial thinking with AI to develop an innovative idea.
• Can show initiative and perseverance in an AI-enabled venture.
• Can reflect on their own entrepreneurial strengths and how AI extends them.

Material on the entrepreneurial mindset with AI — initiative, self-efficacy, taking ideas forward — through projects and reflection.

A startup named as the author of a tool.

• R12 EntreComp Lens for AI Opportunity: R12 applies the EntreComp framework; R04 is the mindset.
• R09 Growth Mindset for AI Era: R09 is continuous learning; R04 entrepreneurial thinking.
• R10 AI & Innovation Management: R10 manages innovation processes; R04 is the individual's mindset.

• Can plan and lead an organisational transition that adopts AI.
• Can address resistance, skills gaps and communication during the change.
• Can measure whether the adoption succeeded.

Material on change management for AI adoption — planning, stakeholders, resistance, training — with cases or plans the learner develops.

A note that 'AI adoption requires change'.

• R08 Digital Transformation Leadership: R08 leads digital transformation; R05 manages a specific transition.
• R17 Adaptation to AI-Driven Change: R17 is the individual adapting; R05 leads the organisation through it.
• R02 Adaptive AI Strategy: R02 sets strategy; R05 carries out the change.

• Can use AI insights to plan recovery after a disruption.
• Can analyse what failed and what to prioritise in recovery.
• Can monitor recovery with data.

Material on post-crisis recovery with AI — analysis of disruption, recovery planning, monitoring — with cases.

IT disaster-recovery settings in a deployment guide.

• R25 Resilient AI Systems Design: R25 makes AI systems resilient; R06 uses AI to recover an organisation.
• R14 Resilience Through Automation: R14 automates to build resilience; R06 recovers after a crisis.
• R03 AI-Informed Risk Assessment: R03 assesses risks before; R06 recovers after.

• Can use AI to create, test and refine a business model.
• Can map a business model with a canvas and test its assumptions with data.
• Can explain how AI changes value creation and capture in a model.

Material on business modelling with AI — canvases, testing assumptions, iterating — with a model the learner builds.

'Business model' used loosely in a product description.

• R16 AI Value Proposition Design: R16 designs the value proposition; R07 the whole business model.
• R01 AI-Enhanced Opportunity Recognition: R01 finds opportunities; R07 builds a model around one.
• R21 Financial Literacy for AI: R21 evaluates costs and returns; R07 designs the model.

• Can lead a team through AI-driven digital transformation.
• Can set a vision, priorities and milestones for the transformation.
• Can develop the team's skills and culture alongside the technology.

Material on leading digital transformation — vision, roadmaps, culture, people — with leadership cases and plans.

A slide titled 'digital transformation' in a vendor deck.

• R05 AI Change Management: R05 manages a specific change; R08 leads the transformation.
• R10 AI & Innovation Management: R10 manages innovation; R08 leads transformation.

• Can keep their knowledge of AI current through deliberate learning.
• Can learn from experience and setbacks with new AI tools.
• Can plan their own reskilling.

Material on the growth mindset and continuous learning in the AI era, with reflection and planning of one's own learning.

An invitation to 'keep learning' at the end of a course.

• R17 Adaptation to AI-Driven Change: R17 adapts professional practice; R09 keeps learning.
• R04 Entrepreneurial Mindset with AI: R04 is entrepreneurial; R09 is about learning.

• Can manage an innovation process that integrates AI, from idea to implementation.
• Can build and prioritise an innovation portfolio.
• Can evaluate innovation outcomes.

Material on innovation management with AI — processes, portfolios, open innovation — with cases or a process the learner designs.

'Innovative' used to describe a product.

• R04 Entrepreneurial Mindset with AI: R04 is the individual's mindset; R10 manages the process.
• D25 AI & Creative Problem Solving: D25 is creative problem solving; R10 organisational innovation.
• R08 Digital Transformation Leadership: R08 leads transformation; R10 manages innovation.

• Can work in a cross-functional team on a problem using AI.
• Can combine the expertise of different roles with AI tools.
• Can facilitate a team's use of AI in problem solving.

Material on collaborative problem solving with AI — team roles, shared tools, facilitation — with team activities.

A contribution guide inviting pull requests.

• D10 Human-AI Collaboration: D10 pairs a person with AI; R11 is a team solving problems with AI.
• R24 Community of Practice Building: R24 builds networks of practice; R11 solves problems in a team.
• R19 Stakeholder Engagement with AI: R19 engages stakeholders; R11 works within a team.

• Can apply EntreComp competences to identify AI-enabled opportunities.
• Can mobilise resources and people for an AI-enabled idea.
• Can assess an idea's value, ethics and sustainability with the framework.

Material that uses EntreComp, or its competences, as a lens on AI opportunities, with activities applying it.

EntreComp cited in a reference list.

• R01 AI-Enhanced Opportunity Recognition: R01 uses AI to find opportunities; R12 applies EntreComp.
• R04 Entrepreneurial Mindset with AI: R04 is the mindset; R12 the framework.
• G13 GreenComp Lens for AI: G13 applies GreenComp; R12 EntreComp.

• Can use AI for market research, competitor analysis and trend detection.
• Can interpret and check AI-generated market insights.
• Can turn the analysis into recommendations.

Material that applies AI to market analysis — data sources, competitor tracking, forecasting — with an analysis the learner carries out.

A market-size figure quoted in an introduction.

• R22 AI-Enabled Customer Insights: R22 analyses customers; R13 markets and competitors.
• R01 AI-Enhanced Opportunity Recognition: R01 spots opportunities; R13 analyses markets.

• Can identify processes whose automation with AI would make an organisation more resilient.
• Can design an automation and its fallback for failures.
• Can assess automation's effect on continuity.

Material on automation for resilience — choosing processes, designing automation with fallbacks, continuity — with cases.

A CI workflow that automates a build.

• D23 AI Workflow Integration: D23 embeds AI into workflows; R14 automates for resilience.
• R26 AI-Powered Process Optimisation: R26 optimises processes; R14 automates them to keep operating.
• R06 Post-Crisis AI Recovery: R06 recovers after a crisis; R14 prepares for one.

• Can apply an ethical framework to a commercial AI deployment.
• Can identify conflicts between commercial goals and ethics.
• Can design safeguards and argue for them in a business setting.

Material on ethics in AI entrepreneurship — frameworks, cases, trade-offs — through business cases and reflection.

A company's code of conduct linked in a footer.

• D22 Responsible AI Use Mindset: D22 is responsible use; R15 ethics in commercial deployment.
• T22 Responsible AI Development: T22 is responsible development; R15 business ethics.
• R23 Regulatory Navigation for AI: R23 navigates regulation; R15 ethics beyond compliance.

• Can design a value proposition that uses AI capabilities.
• Can map customer jobs, pains and gains and match them with AI features.
• Can test a value proposition with customers.

Material on value proposition design with AI — canvases, customer understanding, testing — with a proposition the learner designs.

A product's tagline describing its value.