Source-linked AI summary
Generative AI Literacy: Twelve Defining Competencies
Ravinithesh Annapureddy, Alessandro Fornaroli, Daniel Gatica-Perez
TL;DR
Generative AI literacy is not yet well established in the academic literature, where existing work often conflates it with general AI literacy. This paper addresses that gap by defining twelve competencies for understanding and using generative AI, while noting that the framework will require revision as the technology evolves.
Problem
Generative AI literacy lacks a well-established concept in academic literature, and existing papers often conflate it with general AI literacy.
Method
The paper defines generative AI literacy through twelve competencies covering essential knowledge and skills for understanding and using generative AI systems.
Results
The preliminary search found only 6 records, highlighting the novelty of generative AI literacy and the current gap in academic literature.
Takeaways & Limitations
Comprehending, assessing, and employing generative AI technologies is presented as fundamental to enabling individuals to engage with and contribute to an AI-enabled world.
Takeaways & Limitations
The framework reflects the current state of a rapidly evolving field and will require revisions as generative AI technology advances.
Abstract
from arXiv · showhide
This paper introduces a competency-based model for generative artificial intelligence (AI) literacy covering essential skills and knowledge areas necessary to interact with generative AI. The competencies range from foundational AI literacy to prompt engineering and programming skills, including ethical and legal considerations. These twelve competencies offer a framework for individuals, policymakers, government officials, and educators looking to navigate and take advantage of the potential of generative AI responsibly. Embedding these competencies into educational programs and professional training initiatives can equip individuals to become responsible and informed users and creators of generative AI. The competencies follow a logical progression and serve as a roadmap for individuals seeking to get familiar with generative AI and for researchers and policymakers to develop assessments, educational programs, guidelines, and regulations.
1 INTRODUCTION
The paper proposes a competency-based model for generative AI literacy because existing AI-literacy frameworks do not specifically address generative AI. The model covers understanding, interaction, creation, contextual use, and ethical and practical considerations.
- Generative AI literacy is presented as important for informed adoption, risk assessment, and mitigation in digital government and other professional contexts.
- Generative AI requires distinct literacy because it produces diverse outputs, operates across disciplines, and raises questions about authorship, ownership, originality, and responsible use.
- The paper proposes a competency-based framework for assessing users’ ability to understand, interact with, and create generative AI models.
- Existing AI-literacy frameworks do not differentiate between generative and predictive models, leaving no specific competency set for generative AI use.
- The twelve competencies complement existing AI-literacy frameworks by extending beyond prompt engineering to include technical, contextual, situational, ethical, and practical aspects.
2 METHODOLOGY
The authors conducted a preliminary database search and a non-systematic literature review to identify competencies for a generative AI literacy model. The search found only six unique records, underscoring the topic’s novelty and the scarcity of established literature.
- Preliminary Database Search: A search of five scientific databases found six unique records for “generative AI literacy” after duplicate removal.
- Literature Review: Because the topic was novel and prior research was scarce, the authors used a non-systematic comprehensive literature review rather than a long-standing systematic review.
- Literature Review: The review covered seminal papers and systematic reviews on generative AI applications, AI literacy, prompt engineering, and related competency frameworks.
- Competency Model Development: The authors identified competencies from the literature, then described each competency with an illustrative example and discussed potential implications speculatively.
- State of the Literature: The database search indicated that generative AI literacy was not well established and that existing studies often conflated it with general AI literacy.
4 A COMPETENCY-BASED MODEL FOR GENERATIVE AI LITERACY
The paper defines generative AI literacy as a continuous process involving competencies for understanding, assessing, and working responsibly with generative AI. It proposes twelve competencies arranged as a progressive learning path and outlines their potential positive and negative implications.
- Generative AI literacy is defined as a long-term process of acquiring, practicing, and mastering knowledge, skills, abilities, and behaviors.
- The model identifies twelve competencies needed to comprehend and effectively use generative AI tools.It builds on literature concerning AI literacy, digital literacy, prompt engineering, and generative AI.
- The competencies follow a structured learning path from foundational knowledge and content detection to assessment, prompting, programming, contextual, ethical, and legal understanding.Continuous self-learning is described as transversal across the capabilities.
- Table 3 presents each competency with potential benefits of proficiency and potential risks associated with not acquiring it.The paper states that these implications are mostly speculative because prior research on the concepts is scarce.
1. Basic AI literacy
Basic AI literacy establishes a foundational understanding of AI needed to engage with generative AI.
- Basic AI literacy establishes a baseline understanding of AI for working with generative AI.
2. Knowledge of generative AI models
Knowledge of generative AI models supports understanding their capabilities, constraints, and outputs. It also helps users assess content quality, authenticity, relevance, and usefulness.
- Understanding generative AI models supports assessment of their capabilities and constraints.
- Users should critically assess whether generated content is relevant, useful, accurate, and appropriate for its purpose.
- Model knowledge helps users distinguish generative AI from search engines that retrieve existing information.
- Understanding model behavior sets realistic expectations and can reduce misuse, missed opportunities, or over-reliance.
- Verifying generated content authenticity helps maintain trust in information and AI-supported work.
- The competency also supports creative engagement and informed participation in AI-driven environments.
7. Skill in prompting generative AI tools (Prompt Engineering)
Prompt engineering enables users to tailor generative AI outputs to specific objectives. Programming and contextual knowledge extend this capability by supporting model customization and appropriate application across settings.
- 7. Skill in prompting generative AI tools (Prompt Engineering): Prompt engineering tailors generative AI outputs to users’ specific needs, objectives, and creative efforts.
- 7. Skill in prompting generative AI tools (Prompt Engineering): Prompting skills increase the practical value of generative AI by enabling more personalized outputs.
- 7. Skill in prompting generative AI tools (Prompt Engineering): Limited prompting ability may restrict users to less useful or less tailored outputs.
- 8. Ability to program and fine-tune: Programming and fine-tuning skills support optimization of generative AI models for specific purposes.
- 8. Ability to program and fine-tune: Without programming and fine-tuning ability, users may remain dependent on off-the-shelf solutions.
- 9. Knowledge of the contexts where generative AI is used: Knowledge of generative AI use contexts helps users recognize diverse applications and limitations across situations, institutions, and professions.
- 9. Knowledge of the contexts where generative AI is used: Insufficient contextual knowledge may lead to inappropriate applications of generative AI.
10. Knowledge of the Ethical Implications
The framework treats ethical considerations as part of generative AI use, including potential ethical concerns, societal impacts, and responsibility.
- 10. Knowledge of the Ethical Implications: Ethical considerations are tied directly to the use of generative AI.
- 10. Knowledge of the Ethical Implications: Responsible use involves developing a sense of responsibility toward generative AI’s implications.
- 10. Knowledge of the Ethical Implications: The framework highlights potential ethical concerns and societal impacts associated with generative AI.
5 GENERATIVE AI COMPETENCIES
The competency model progresses from general AI literacy and understanding generative AI to practical tool use, evaluation, prompting, programming, and responsible contextual application.
- Foundational knowledge: Foundational competencies require understanding AI systems and generative AI models, including their statistical basis, training data, outputs, capabilities, and limitations.
- Tool knowledge and use: Users should select appropriate generative tools, learn emerging models, and match tools to specific applications.
- Evaluation and assessment: Generative AI literacy includes detecting AI-generated content and critically assessing outputs using trusted sources to reduce hallucination effects.
- Advanced and responsible use: Advanced competencies include prompt engineering, programming-related model development, and adapting generative AI use to social, professional, ethical, legal, and environmental contexts.
- Legal knowledge and continuous learning: Users should understand applicable legal frameworks and continuously learn new tools, issues, and regulations.
6 DISCUSSION
The discussion presents the twelve competencies as a broad framework for responsible generative AI use, education, assessment, and governance. It also emphasizes that the framework must evolve with the technology and may require domain-specific extensions.
- 6.1 Advantages and Limitations: The competencies move users from consumers to interpreters who can optimize generative AI for problem-solving and engage with generated content critically.The framework positions AI as a collaborator and supports active participation in its development, application, and ethical use.
- 6.2 Applications: The framework supports educational curricula, professional development, and outreach programs for learners and workers at different levels.It can guide curricula from elementary through higher education, workplace training, and inclusive community initiatives.
- 6.1 Advantages and Limitations: Because generative AI spans domains and evolves rapidly, the framework may need domain-specific sub-skills and continuing revisions.The discussion also identifies future research on long-term decision-making effects and human–AI collaboration involving communication, trust, and symbiotic relationships.
- 6.2 Applications: Public institutions can use the competencies to assess employee capabilities, identify adoption needs, and develop training and responsible-use policies.Policymakers can also use them to inform guidelines addressing ethical, legal, political, accountability, transparency, risks, and biases.
- 6.3 Future Research Directions: The competency-based model provides a basis for assessments that define tasks and metrics, classify proficiency levels, and create teaching resources.These assessments can be used in educational settings, certification programs, and workplaces to evaluate conceptual, capability, and ethical understanding.
7 CONCLUSION
The conclusion defines generative AI literacy through twelve competencies spanning understanding, use, creation, adaptive practice, ethics, and law. It presents the framework as a starting point for education, training, and broader standardization that must be revised as AI evolves.
- 7 CONCLUSION: The paper identifies twelve competencies that combine knowledge and skills for understanding, assessing, and using generative AI systems.The competencies extend beyond technical proficiency to adaptive usage, ethical considerations, and legal awareness for a broad audience.
- 7 CONCLUSION: The competencies aim to help users develop realistic expectations, avoid overreliance, and use AI as a collaborator or tutor within their domains.The conclusion frames these capacities as supporting effective engagement with AI technologies rather than passive consumption.
- 7 CONCLUSION: The framework is a starting point rather than a static standard, requiring revision as AI technologies and applications evolve.The paper describes generative AI literacy as evolving and calls for continued updates to the competencies.
- 7 CONCLUSION: Embedding the competencies in curricula, training, and professional development can support informed participation in AI-driven initiatives.The conclusion presents coordinated implementation as groundwork for a potential global standardization of generative AI literacy.