These appendices present the full competency rubrics used in the proposed framework for AI competencies in engineering education. They are intended to support curriculum design, faculty development, self-assessment, and program evaluation.
Appendix A. AI Competency Checklist for Any Engineer
This checklist summarizes the AI competencies that any engineer should develop, regardless of specialization. The expected level is competent performance in core AI literacy, responsible use, engineering applications, and interdisciplinary communication.
| Domain | Checklist indicator |
|---|---|
| AI foundations | Distinguishes symbolic AI, machine learning, generative AI, and agent-based AI |
| Historical evolution | Explains the transition from rules and search to ML, LLMs, and agents |
| Non-generative AI | Recognizes search, rules, Bayesian reasoning, planning, and optimization |
| Search and problem solving | Represents problems as states, actions, paths, costs, and goals |
| Constraints and assignment | Identifies variables, domains, constraints, scheduling, and allocation problems |
| Logical and probabilistic reasoning | Interprets rule-based and uncertainty-based reasoning |
| Planning and decision-making | Recognizes actions, goals, costs, constraints, and decision sequences |
| Optimization | Connects optimization with routing, design, control, and resource allocation |
| Machine learning | Distinguishes supervised, unsupervised, and reinforcement learning |
| Neural networks | Recognizes common neural architectures and their applications |
| LLMs and foundation models | Understands tokens, embeddings, attention, prompting, and limitations |
| Intelligent agents | Identifies environment, state, action, goal, memory, tools, and feedback |
| Multi-agent systems | Recognizes coordination, cooperation, competition, and emergent behavior |
| Engineering applications | Connects AI to robotics, manufacturing, logistics, energy, software, maintenance, and biotechnology |
| Responsible use | Evaluates risks related to bias, privacy, hallucinations, safety, and accountability |
| Communication | Explains AI assumptions, outputs, risks, and limitations to interdisciplinary teams |
Appendix B. AI Competency Checklist for Computing-Related Engineers and Data Scientists
This checklist summarizes specialized AI competencies for profiles responsible for designing, implementing, evaluating, deploying, and governing AI systems. The expected level is competent to advanced performance in technical AI development.
| Domain | Checklist indicator |
|---|---|
| Mathematical foundations | Applies linear algebra, calculus, probability, statistics, optimization, and uncertainty |
| Classical AI | Formalizes problems using states, actions, rules, constraints, probabilities, and plans |
| Search and planning | Implements and evaluates BFS, DFS, A*, heuristic search, and planning strategies |
| CSP and combinatorial optimization | Solves assignment, scheduling, configuration, and constrained design problems |
| Nature-inspired algorithms | Applies GA, PSO, ACO, simulated annealing, or hybrid metaheuristics |
| Data pipelines | Cleans, transforms, partitions, documents, and monitors data |
| Supervised learning | Trains, validates, tunes, and interprets predictive models |
| Unsupervised learning | Applies clustering, dimensionality reduction, embeddings, and anomaly detection |
| Semi/self-supervised learning | Uses unlabeled or partially labeled data through pretraining or pseudo-labeling |
| Reinforcement learning | Models states, actions, rewards, policies, and sequential decisions |
| Transfer and lifelong learning | Applies fine-tuning, domain adaptation, continual learning, or few-shot strategies |
| Federated learning and privacy | Designs privacy-aware learning with distributed data |
| Neural architectures | Selects or implements architectures for images, text, signals, graphs, or multimodal data |
| Transformers and LLMs | Uses prompting, RAG, fine-tuning, semantic search, extraction, summarization, and evaluation |
| Tool-using agents | Designs agents with memory, tools, APIs, retrieval, code execution, and supervision |
| Multi-agent simulation | Models coordination, roles, communication, negotiation, and emergent behavior |
| Model evaluation | Uses validation, metrics, error analysis, robustness, fairness, and out-of-sample testing |
| Explainability | Applies SHAP, LIME, sensitivity analysis, saliency maps, or interpretable models |
| Software engineering for AI | Uses version control, testing, modularity, reproducibility, documentation, and pipelines |
| MLOps | Deploys, monitors, retrains, secures, and maintains AI systems |
| Ethics and governance | Applies privacy, fairness, traceability, human oversight, and accountability |
| Technical communication | Translates domain problems into data/modeling problems and explains limitations |
Appendix C. AI Competency Checklist for Engineering Educators
This checklist summarizes the pedagogical AI competencies required by engineering educators. The expected level is competent to advanced performance in pedagogical integration, curriculum redesign, authentic assessment, and responsible AI use.
| Domain | Checklist indicator |
|---|---|
| AI literacy for teaching | Explains AI concepts using examples from engineering education |
| AI-ready STEAM foundations | Connects linear algebra, probability, statistics, optimization, programming, and modeling with AI concepts |
| Curriculum redesign | Aligns AI competencies with learning outcomes, activities, evidence, and assessment |
| Generative AI for instructional design | Uses GenAI to create examples, cases, simulations, feedback, rubrics, and activities |
| AI-assisted active learning | Designs PBL, projects, simulations, escape rooms, Socratic dialogues, and educational games |
| Teaching symbolic AI | Designs activities involving search, constraints, logic, planning, and optimization |
| Teaching data-driven AI | Uses datasets, visualizations, model evaluation, uncertainty, and error analysis |
| Teaching with LLMs | Guides prompting, verification, comparison, documentation, and critical use |
| Teaching agents and multi-agent systems | Uses agents, simulations, games, robots, or human-AI collaboration scenarios |
| Authentic assessment | Requires justification, validation, comparison, critique, and reflection |
| Detection of superficial learning | Uses oral defense, process evidence, logs, critique tasks, and in-class validation |
| Metacognition | Helps students question, verify, iterate, self-explain, and make responsible decisions |
| Ethical use | Defines norms for disclosure, privacy, academic integrity, fairness, and accountability |
| Inclusion and equity | Uses AI to support differentiated, accessible, multilingual, and personalized learning |
| Evidence-based improvement | Collects evidence from student work, feedback, analytics, rubrics, and observations |
| Professional development | Updates AI foundations, tools, risks, pedagogical strategies, and institutional practices |
Appendix D. AI Competency Rubrics for Engineering Education
AI Knowledge and Skills Every Engineer Should Have
This rubric assesses the level of understanding, application, and critical judgment that every engineer should develop regarding artificial intelligence, regardless of specialization. It focuses on the ability to understand fundamental concepts, identify engineering applications, use AI tools responsibly, and collaborate with technical specialists.
| Criterion | Initial | Basic | Competent | Advanced |
|---|---|---|---|---|
| 1. General understanding of AI | Recognizes AI as a current technology, but associates it mainly with chatbots or automation. | Distinguishes, in general terms, among traditional AI, machine learning, and generative AI. | Explains AI as a set of techniques for search, reasoning, learning, optimization, and content generation. | Historically relates symbolic, connectionist, statistical, generative, and agent-based AI, identifying continuities and differences. |
| 2. History and evolution of AI | Identifies some isolated AI milestones. | Recognizes general stages: symbolic AI, expert systems, neural networks, machine learning, and generative AI. | Explains how AI has evolved from rules and search to foundation models and agents. | Critically analyzes hype cycles, technical limitations, computational advances, and changes in data, hardware, and algorithms. |
| 3. Non-generative AI | Has a general idea that some AI algorithms do not generate text or images. | Recognizes techniques such as search, rules, Bayesian networks, optimization, and supervised learning. | Explains applications of search, constraint satisfaction, reasoning, planning, and optimization in engineering problems. | Models real engineering problems as search, constraints, planning, probabilistic inference, or optimization. |
| 4. Search algorithms and problem solving | Recognizes that some problems can be represented as paths or states. | Distinguishes between uninformed and informed search. | Explains algorithms such as BFS, DFS, uniform-cost search, and A*, relating them to routes, graphs, robotics, or logistics. | Selects an appropriate search strategy considering cost, optimality, completeness, and complexity. |
| 5. Constraint satisfaction and assignment | Identifies that many problems have conditions that must be satisfied. | Recognizes examples such as schedules, resource allocation, or system configuration. | Formulates problems with variables, domains, and constraints. | Differentiates among constraint satisfaction, constrained optimization, assignment, scheduling, and planning. |
| 6. Logical and probabilistic reasoning | Recognizes that AI can use rules or probabilities. | Distinguishes between rule-based reasoning and reasoning under uncertainty. | Explains expert systems, inference rules, Bayesian networks, and probabilistic diagnosis. | Interprets causal or probabilistic relationships in contexts involving risk, sensors, failures, maintenance, and safety. |
| 7. Planning and decision-making | Identifies that an intelligent system can follow a sequence of actions. | Recognizes examples of planning in robotics, manufacturing, or logistics. | Describes planning as constructing a sequence of actions from an initial state toward a goal. | Analyzes problems involving actions, constraints, uncertainty, costs, and multiple objectives. |
| 8. Optimization and nature-inspired algorithms | Recognizes that AI can search for better solutions. | Identifies algorithms such as genetic algorithms, particle swarm optimization, ant colony optimization, or simulated annealing. | Relates nature-inspired algorithms to design, routing, assignment, control, and scheduling. | Selects optimization approaches according to problem type: continuous, discrete, combinatorial, multiobjective, or constrained. |
| 9. Machine learning | Recognizes that models learn from data. | Distinguishes supervised, unsupervised, and reinforcement learning. | Explains supervised, unsupervised, semi-supervised, self-supervised, reinforcement, transfer, federated, lifelong, and meta-learning. | Evaluates which learning paradigm is appropriate according to data availability, labels, privacy, environment, and need for adaptation. |
| 10. Neural architectures and topologies | Recognizes terms such as neural network or deep learning. | Identifies some architectures: MLP, CNN, RNN, Transformer. | Distinguishes topology, architecture, model, and learning algorithm. | Relates architectures to data types: images, signals, text, graphs, time series, and multimodal data. |
| 11. LLMs and foundation models | Uses LLMs as tools without understanding their basic operation. | Recognizes that LLMs work with tokens, embeddings, and text prediction. | Explains the relationship among Transformers, attention, embeddings, self-supervised learning, and generation. | Distinguishes generative and non-generative uses of LLMs: classification, retrieval, analysis, agents, programming, and assisted decision-making. |
| 12. AI agents | Identifies an agent as a system that responds to instructions. | Recognizes that an agent can perceive, decide, and act. | Explains components such as environment, state, action, goal, memory, tools, and evaluation. | Conceptually designs agents to solve engineering problems using search, planning, learning, tools, or LLMs. |
| 13. Multi-agent systems | Recognizes that a system can have multiple autonomous components. | Identifies examples such as robots, vehicles, sensors, users, or interacting machines. | Explains cooperation, competition, coordination, communication, and emergent behavior in engineering problems. | Analyzes when it is appropriate to model a problem as a multi-agent system and what risks it involves: conflicts, scalability, uncertainty, and control. |
| 14. Engineering applications | Mentions general AI applications. | Relates AI to automation, prediction, classification, and optimization. | Identifies applications by area: robotics, manufacturing, energy, civil engineering, biomedical engineering, logistics, software, and maintenance. | Proposes viable AI solutions considering data, objectives, constraints, risks, cost, impact, and scalability. |
| 15. Responsible and ethical use | Recognizes that AI has risks. | Identifies issues such as bias, privacy, errors, and excessive dependence. | Evaluates risks related to hallucinations, sensitive data, explainability, security, responsibility, and equity. | Integrates ethical, legal, technical, and social criteria in the selection, implementation, and evaluation of AI solutions. |
| 16. Interdisciplinary communication and collaboration | Uses AI terms imprecisely. | Communicates basic AI ideas to other engineers. | Clearly explains problems, models, data, assumptions, and limitations. | Collaborates effectively with specialists in data, computing, engineering domains, operations, management, and end users. |
Suggested interpretation
| Level | Descriptor |
|---|---|
| Initial | Knows AI mainly as a user. |
| Basic | Understands fundamental concepts and recognizes applications. |
| Competent | Can analyze engineering problems and associate them with AI techniques. |
| Advanced | Can participate in the design, evaluation, and decision-making of AI solutions in real contexts. |
Expected profile for any engineer
Every engineer should reach at least a competent level in general understanding, applications, responsible use, problem identification, and interdisciplinary collaboration. They do not necessarily need to program advanced models, but they should be able to communicate with specialists, evaluate relevance, interpret results, and recognize limitations.
AI Knowledge and Skills for Computing Engineers and Data Scientists
This rubric assesses more specialized competencies for profiles responsible for designing, implementing, training, evaluating, deploying, and maintaining artificial intelligence systems. It is intended for computing engineers, data scientists, machine learning specialists, AI developers, and related technical profiles.
| Criterion | Initial | Basic | Competent | Advanced |
|---|---|---|---|---|
| 1. Mathematical foundations for AI | Handles isolated notions of algebra, calculus, probability, or statistics. | Uses basic concepts of vectors, matrices, derivatives, distributions, and error. | Applies linear algebra, optimization, probability, statistical inference, and calculus to understand AI models. | Formulates, analyzes, and justifies models using mathematical foundations, complexity, convergence, regularization, and uncertainty. |
| 2. Representation of classical AI problems | Recognizes search or rule-based problems, but does not formalize them. | Models simple problems with states, actions, and goals. | Implements uninformed search, informed search, CSP, logical reasoning, Bayesian networks, and basic planning. | Designs hybrid solutions combining search, logic, probability, planning, and optimization in complex problems. |
| 3. Search and planning algorithms | Knows names of algorithms such as BFS, DFS, or A*. | Implements simple versions of BFS, DFS, and A*. | Evaluates algorithms according to completeness, optimality, memory, time, and heuristic quality. | Designs heuristics, planners, or adapted variants for robotics, video games, logistics, automation, or multi-agent systems. |
| 4. Constraint satisfaction and combinatorial optimization | Identifies assignment or constraint problems. | Formulates simple CSPs with variables, domains, and constraints. | Implements backtracking, constraint propagation, local search, and constrained optimization models. | Solves real scheduling, assignment, design, or configuration problems using CSP, integer programming, metaheuristics, or hybrid approaches. |
| 5. Nature-inspired algorithms | Recognizes examples such as genetic algorithms or swarms. | Explains the inspiration and basic operation of GA, PSO, ACO, or simulated annealing. | Implements and tunes metaheuristics for optimization problems. | Compares metaheuristics, designs operators, and evaluates convergence, diversity, exploration-exploitation, and multiobjective performance. |
| 6. Data preparation and understanding | Loads data and performs basic inspections. | Cleans data and identifies missing values, outliers, and relevant variables. | Builds data pipelines with validation, transformation, partitioning, encoding, scaling, and documentation. | Designs robust data quality, lineage, governance, privacy, bias, and distribution-shift monitoring strategies. |
| 7. Supervised learning | Trains basic models without deep evaluation. | Applies regression, classification, and common metrics. | Selects models, tunes hyperparameters, validates results, and controls overfitting. | Designs robust supervised solutions considering imbalance, interpretability, uncertainty, error cost, and generalization. |
| 8. Unsupervised learning and representation | Applies clustering or dimensionality reduction in a basic way. | Uses k-means, PCA, or other exploratory methods. | Interprets clusters, embeddings, dimensionality reduction, anomalies, and latent structure. | Designs useful representations for semantic search, anomaly detection, segmentation, recommendation, or pretraining. |
| 9. Semi-supervised and self-supervised learning | Recognizes that labels are not always available. | Distinguishes semi-supervised from self-supervised learning. | Applies techniques to leverage unlabeled or partially labeled data. | Designs pretext tasks, pseudo-labeling strategies, contrastive learning, or pretraining for specific domains. |
| 10. Reinforcement learning | Recognizes agent, environment, action, and reward. | Implements simple examples of Q-learning or basic policies. | Formulates MDPs, defines rewards, evaluates policies, and trains agents in simulation. | Designs RL systems for control, robotics, sequential optimization, or decision-making under uncertainty, considering safety and stability. |
| 11. Transfer learning, lifelong learning, and meta-learning | Recognizes that a model can reuse knowledge. | Applies pretrained models to simple tasks. | Uses fine-tuning, domain adaptation, and strategies for limited data. | Designs adaptive systems with transfer, continual learning, catastrophic forgetting prevention, and rapid adaptation to new tasks. |
| 12. Federated learning and privacy | Recognizes that data can be distributed. | Explains the idea of training without centralizing data. | Implements or designs basic federated learning schemes. | Integrates differential privacy, security, robust aggregation, data heterogeneity, and regulatory constraints. |
| 13. Neural architectures | Uses neural networks as black boxes. | Distinguishes MLP, CNN, RNN, LSTM, Autoencoder, and Transformer. | Implements, trains, and evaluates architectures according to data type: image, text, signal, graph, or time series. | Designs architectures or adapts advanced models considering capacity, regularization, efficiency, interpretability, and computational cost. |
| 14. Transformers, LLMs, and foundation models | Uses LLMs without understanding their architecture. | Explains tokens, embeddings, attention, and token prediction. | Applies LLMs for classification, extraction, summarization, generation, semantic search, RAG, and programming assistance. | Designs evaluable, safe, and traceable LLM systems using advanced prompting, RAG, fine-tuning, agents, tools, and systematic evaluation. |
| 15. Tool-using agents and multi-agent systems | Recognizes agents as automated assistants. | Designs simple workflows with input, reasoning, and output. | Builds agents that use memory, tools, APIs, databases, code execution, or information retrieval. | Designs multi-agent architectures with planning, coordination, evaluation, safety, traceability, action limits, and human supervision. |
| 16. Multi-agent systems and agent-based simulation | Recognizes multiple agents interacting in an environment. | Implements simple simulations with local rules. | Designs multi-agent systems with communication, coordination, roles, objectives, and performance metrics. | Implements complex systems with multi-agent learning, negotiation, game theory, distributed coordination, emergent behavior, and robust evaluation. |
| 17. Model evaluation | Uses accuracy or error without further analysis. | Applies basic classification, regression, or clustering metrics. | Evaluates with appropriate metrics, cross-validation, ROC/PR curves, confusion matrices, error, bias-variance, and out-of-sample tests. | Designs comprehensive evaluation frameworks with technical metrics, business cost, robustness, fairness, uncertainty, safety, and post-deployment monitoring. |
| 18. Interpretability and explainability | Recognizes that some models are difficult to interpret. | Uses variable importance or simple explanations. | Applies techniques such as SHAP, LIME, sensitivity analysis, saliency maps, or tree interpretability. | Integrates explainability into critical systems, audits, human decision-making, and communication with non-technical users. |
| 19. Software engineering for AI | Writes functional but poorly maintainable scripts. | Organizes code, data, and experiments in a basic way. | Uses version control, testing, reproducible environments, documentation, and pipelines. | Designs robust ML systems with modularity, tests, CI/CD, data/model versioning, monitoring, and scalability. |
| 20. MLOps and deployment | Trains models locally without deploying them. | Exports models and uses them in simple applications. | Deploys models as services, APIs, dashboards, or automated processes. | Manages the full lifecycle: training, validation, deployment, monitoring, retraining, drift, security, and costs. |
| 21. AI ethics, security, and governance | Recognizes general AI risks. | Identifies bias, privacy, hallucinations, errors, and misuse. | Applies principles of security, privacy, traceability, evaluation, and human oversight. | Designs governance, audit, risk mitigation, regulatory compliance, and accountability frameworks for AI systems. |
| 22. Technical communication and domain translation | Explains models with overly technical or superficial language. | Communicates basic results with graphs and metrics. | Translates domain problems into data/modeling problems and communicates assumptions, results, and limitations. | Leads conversations among business, engineering, data, operations, and users, making technically and ethically justified decisions. |
Suggested interpretation
| Level | Descriptor |
|---|---|
| Initial | Uses AI tools or basic implementations, but with limited understanding. |
| Basic | Can apply known models following standard procedures. |
| Competent | Can design, train, evaluate, and deploy AI solutions for real problems. |
| Advanced | Can create robust, scalable, auditable solutions adapted to complex contexts. |
Expected profile for computing engineers and data scientists
A computing engineer or data scientist should reach at least a competent level in mathematical foundations, programming, data, models, evaluation, deployment, and ethics. To lead AI projects, they should approach the advanced level in system design.
AI Competencies for Engineering Educators
This rubric assesses the pedagogical AI competencies that engineering educators should develop to understand AI foundations, integrate generative AI critically, redesign curricula, create meaningful learning experiences, evaluate student learning authentically, and guide responsible human-AI collaboration.
| Criterion | Initial | Basic | Competent | Advanced |
|---|---|---|---|---|
| 1. AI literacy for teaching | Recognizes AI mainly as a classroom tool or chatbot. | Explains basic differences among symbolic AI, machine learning, generative AI, and agents. | Uses engineering examples to explain AI concepts, applications, limits, and risks. | Integrates historical, conceptual, technical, ethical, and pedagogical perspectives to teach AI critically. |
| 2. AI-ready STEAM foundations | Mentions mathematics or computing as generally related to AI. | Identifies links between AI and linear algebra, probability, statistics, optimization, programming, or modeling. | Connects foundational STEAM topics to embeddings, uncertainty, learning, search, simulation, and model evaluation. | Redesigns early STEAM learning pathways so students build foundations for modern AI theory and practice. |
| 3. Curriculum redesign | Adds AI tools to existing activities without changing learning outcomes. | Includes AI-related topics or activities in selected course components. | Aligns AI competencies with learning outcomes, activities, evidence, assessment, and ethical use. | Leads flexible, competency-based curriculum redesign that integrates AI progressively across courses or programs. |
| 4. Generative AI for instructional design | Uses GenAI mainly to produce texts, slides, or simple examples. | Uses GenAI to draft materials, examples, questions, or rubrics with teacher review. | Designs AI-assisted cases, simulations, feedback, rubrics, and activities aligned with learning goals. | Creates robust instructional design workflows with GenAI, validation criteria, disciplinary accuracy, and continuous improvement. |
| 5. AI-assisted active learning | Uses AI as an optional support for individual student work. | Introduces AI into guided activities such as discussions, practice, or problem solving. | Designs PBL, projects, simulations, escape rooms, Socratic dialogues, or educational games with AI support. | Orchestrates complex active learning environments where AI supports inquiry, collaboration, creativity, and deep reasoning. |
| 6. Teaching symbolic and non-generative AI | Mentions classical AI topics without integrating them into activities. | Explains basic search, rules, logic, planning, constraints, or optimization examples. | Designs activities where students model engineering problems using search, constraints, reasoning, planning, or optimization. | Connects symbolic AI with modern AI, agents, simulation, and engineering decision-making in integrated learning experiences. |
| 7. Teaching data-driven AI | Uses datasets or model outputs without deep analysis. | Introduces data, prediction, classification, metrics, or basic model evaluation. | Guides students through data preparation, visualization, learning algorithms, error analysis, and uncertainty. | Designs evidence-rich activities where students critique data quality, bias, model limits, and engineering implications. |
| 8. Teaching with LLMs and generative AI | Allows LLM use without clear guidance or verification practices. | Defines basic rules for prompting, citation, verification, or disclosure. | Guides students to compare, validate, improve, and document AI-generated outputs. | Designs advanced human-AI collaboration tasks involving prompting, RAG, critique, iteration, and transparent reasoning. |
| 9. Teaching agents and multi-agent systems | Mentions agents as AI assistants or automated systems. | Uses simple examples of agents that perceive, decide, or act. | Designs activities with agents, simulations, games, robots, or human-AI collaboration scenarios. | Creates multi-agent learning environments to study coordination, emergence, autonomy, supervision, and ethical risks. |
| 10. Authentic assessment | Uses traditional product-based assessments that may be easily automated. | Adds questions about AI use or asks students to disclose tool use. | Designs assessments requiring justification, validation, comparison, critique, reflection, and oral or process evidence. | Develops assessment systems that evaluate reasoning, process, collaboration, ethical judgment, and AI-supported performance. |
| 11. Detection of superficial learning | Notices possible overdependence on AI but lacks clear evidence strategies. | Uses simple checks such as oral questions, in-class work, or plagiarism-style review. | Uses process logs, prompt records, version histories, critique tasks, and validation activities to identify shallow understanding. | Builds assessment ecosystems that make student reasoning visible and reduce uncritical delegation to AI. |
| 12. Metacognition and human-AI collaboration | Encourages students to use AI but gives limited guidance on thinking processes. | Prompts students to review AI outputs and explain whether they agree. | Teaches students to question, verify, iterate, self-explain, and make responsible decisions with AI. | Cultivates reflective human-AI collaboration where students use AI to deepen reasoning, not replace it. |
| 13. Ethical and transparent AI use | Recognizes general concerns about academic integrity or misuse. | Defines basic classroom norms for disclosure, privacy, and acceptable AI use. | Integrates bias, fairness, privacy, transparency, accountability, and academic integrity into learning activities. | Leads ethical AI policies and practices aligned with institutional governance, disciplinary standards, and student agency. |
| 14. Inclusion and educational equity | Recognizes that AI may help or harm different learners. | Uses AI to provide additional explanations, practice, or language support. | Designs differentiated, accessible, multilingual, and personalized learning supports with human oversight. | Evaluates the equity impact of AI-supported teaching and adapts practices to reduce barriers and promote inclusion. |
| 15. Evidence-based improvement | Relies mainly on impressions about whether AI improves learning. | Collects basic feedback from students about AI-supported activities. | Uses student work, rubrics, feedback, analytics, and observations to refine AI-assisted teaching. | Conducts systematic inquiry or scholarship of teaching to evaluate and improve AI-supported learning designs. |
| 16. Continuous professional development | Explores AI tools sporadically and informally. | Participates in basic training on AI tools or responsible use. | Updates knowledge of AI foundations, tools, risks, pedagogy, and assessment through structured professional development. | Leads communities of practice, faculty training, curriculum innovation, and institutional capacity building for AI education. |
Suggested interpretation
| Level | Descriptor |
|---|---|
| Initial | Uses or recognizes AI tools with limited pedagogical, technical, or ethical integration. |
| Basic | Explains basic AI ideas and incorporates AI into selected teaching tasks with guidance. |
| Competent | Designs AI-supported learning experiences, assessments, and ethical use guidelines aligned with engineering education goals. |
| Advanced | Leads curriculum redesign, faculty development, evidence-based improvement, and responsible pedagogical innovation with AI. |
Expected profile for engineering educators
An engineering educator should reach at least a competent level in AI literacy for teaching, curriculum redesign, generative AI-supported learning, authentic assessment, ethical use, and continuous professional development. Teachers who lead institutional innovation should approach the advanced level in curriculum transformation, evidence-based improvement, faculty training, and responsible AI governance.