AI Literacy Defined

Before 2021, the prevalence of AI Literacy within the scope of higher education research was relatively modest. An exploration of literature via Google Scholar using the terms "AI Literacy" and "Higher Education" yielded merely 471 articles for the period spanning 2020 to 2021. However, the advent of ChatGPT as an open-source program marked a pivotal expansion in the field, with an increase to over 3,250 articles listed in a similar Google Scholar search for the year 2023. This increase in scholarly articles parallels the rapid technological advancements in AI, underscoring a heightened scholarly interest and recognition of the importance of AI literacy. 

AI literacy has evolved into a critical competency for students across all academic fields. Mastery of AI involves a deep understanding of its applications and potential societal implications, which is essential for proficient navigation of the digital realm and for creating citizens who are capable of engaging critically with the technologies that shape their lives. AI literacy empowers students to critically assess AI technologies, comprehend their foundational mechanisms, and predict their impacts on society. By embedding AI literacy into academic courses, educators equip students with essential skills for success in a world increasingly dominated by AI, transforming them from passive recipients of technological advancements into active, informed contributors to the technological landscape.

For more evidence of how AI literacy impacts student learning, see the following resources:

• AI Literacy Defined

  • Conceptualizing AI literacy: An exploratory review, 2021, Tsz Kit Ng et al. 

  • AI Literacy in Higher and Adult Education: A Scoping Literature Review, 2022, Laupichler et al. 

  • Measuring user competence in using AI, 2022, Wang et al. 

  • Define, foster, and assess student and teacher AI Literacy and competency for all, 2024, Chiu and Sanusi

• AI Literacy Research 

  • A systematic review of AI literacy conceptualization, constructs, and implementation and assessment efforts (2019–2023), 2024, Almatrafi et al. 

  • Embracing the future of Artificial Intelligence in the classroom: the relevance of AI literacy, prompt engineering, and critical thinking in modern education, 2024, Walter

  • Developing a model for AI Across the Curriculum, 2023, Southworth et al.

  • Combining human and artificial intelligence for enhanced AI literacy in higher education, 2024, Tzirides et al.

  • The Human-Centred Design of a Universal Module for Artificial Intelligence Literacy in Tertiary Education Institutions, 2024, De Silva et al. 

AI Literacy and Your Learning Objectives

Bloom's taxonomy and AI 

Bloom's Taxonomy provides a hierarchical framework for classifying learning objectives. As AI tools become increasingly prevalent in educational settings, faculty must reconsider how these technologies interact with and potentially reshape the cognitive processes outlined in Bloom's framework. The integration of AI in classrooms challenges traditional notions of knowledge acquisition and application, particularly at the lower levels of the taxonomy. This shift necessitates a reevaluation of learning objectives to ensure they remain relevant and challenging in an AI-assisted environment. Faculty can consider designing assessments and activities that leverage AI capabilities while still fostering higher-order thinking skills such as analysis, evaluation, and creation. The judicious application of Bloom's Taxonomy in this context can guide educators in developing learning experiences that prepare students for a world where AI augments human cognition. 

For more information on Bloom's taxonomy and AI course integration see the following resources:

• Bloom's Taxonomy in the Interaction Between Artificial Intelligence and Human Learning (2023)

• Advancing meaningful learning in the age of AI (2023)

• The trouble with Bloom’s taxonomy in an age of AI (2024)

• How can generative AI intersect with Bloom’s taxonomy? (2023)

Incorporating AI Literacy Into Your Courses

When designing a course that either integrates AI extensively or employs AI to facilitate learning without it being the central focus, it is crucial to consider AI Literacy and its core competencies. This consideration is essential irrespective of the degree of AI utilization in the course. Educators and students alike must be well-versed in the following AI Literacy competencies before engaging with AI tools in an educational setting. This foundational knowledge ensures that all participants are prepared to navigate the complexities of AI usage effectively and responsibly. 

Core Competencies in AI Literacy - Without AI being the Central Focus of Your Course (Link)

When AI is utilized as a supplementary tool in courses where it is not the primary focus, students must achieve a comprehensive understanding of how to employ AI effectively. This requirement aligns with the instructional approach adopted for teaching other forms of literacy, such as media, social, data, research, or digital literacy. To adeptly navigate and utilize AI, students are required to master specific competencies, mirroring the foundational skills imparted in these other literacy domains.

1. Recognize - awareness of where AI is being used

2. Know and Understand - basic skills and knowledge of why we use AI

3. Use and Apply - how to operate AI Tools to accomplish a task

4. Evaluate - interpret outcomes and how to best utilize AI-created works

5. Create* - design and code AI applications 

6. Navigate Ethically - recognize and understand the ethical and societal implications of AI technology

*not as widely used in classrooms outside enginerring or computer science. 

Using AI in the Classroom - It's more than a one-off Assignment. 

Using AI appropriately incorporates a set of skills (see above), that cannot be taught in one-off assignments. Just like other literacies, it takes time, practice, and patience to fully embrace and use this technology appropriately. Below is a model for using AI in the classroom. This model is a multi-day endeavor where students start with guided practice and play with an AI Tool, then move on to research and discuss the AI ethical considerations, and lastly learn best practices for using AI to achieve the course or assignment goal. For more information on Ethical Considerations see this page and see below for more on Prompting and Evaluating.  

Designing a Course Focused on AI - Topics to Consider (Link) 

When designing a course where AI is the central focus, research suggests tailoring the following topics to the subject/discipline being taught. 

• Introduction to AI - History, Evolution, Definition

• Understanding Machine Learning - Large Language Models

• Proper Data Handling - How AI is trained with data

• AI in Practice - Real-world Applications

• Human AI Interaction - working with AI

• AI and Creativity - Art, Music, Writing

• Critical Thinking - Assess news, research and claims, hype vs. reality

• AI Governance and Policy - Safety, Standards, International Perspectives

• Future Trends and Research - Learn and Predict

• Hands-on Experience - Use as a tool in real or simulated projects

• Ethical AI - Use ethically, responsibly, and sustainably. 

• AI Literacy for All - Discussing AI accessibility and diversity

• Teaching Prompt Engineering - Best Practices

AI Across the Curriculum Model

Southworth et al (2023) from Southern Florida University suggests an "AI Across the Curriculum Model". This model includes the following 6 learning outcomes for AI that can be tailored or applied to specific courses. 

1. Identify, describe, and/or explain the components, requirements and/or characteristics of AI.

2. Recognize, identify, describe, define, and/or explain applications of AI in multiple domains.

3. Select and/or utilize AI tools and techniques appropriate to a specific context and application.

4. Develop, apply, and/or evaluate contextually appropriate ethical frameworks to use within AI.

5. Assess the context-specific value or quality of AI tools and applications.

6. Conceptualize and/or develop tools, hardware, data, and/or algorithms utilized in AI solutions.

Creating Transparent AI-Centric Assignments

When creating assignments where students are required to use AI it is important to decide what role the AI will play in the assignment. Setting clear expectations and transparent guidelines for how students should use and cite AI work will leave no room for interpretation. Creating an AI use statement for every assignment, explicitly outlining how students should use the AI tool, will work as a guide for how students can use AI appropriately and not violate academic integrity. For more information on Transparent Assignment Design see these resources: 

• The Ohio State University - AI Teaching Strategies: Transparent Assignment Design 

• Cal State - Transparency in Learning and Teaching (TILT)

• Cornell University - AI in Assignment Design 

• Duke - Generative AI Assignments 

• Northwestern - Generative AI and Transparent Design in Research Assignments 

How Can Your Students Use AI

Before deciding on a role for AI, it is imperative that educators first understand the myriad of ways AI can be helpful to their students. Below is a table with examples of how students can use AI as a writing partner, a study buddy, and a co-researcher.

Additionally, below is a more comprehensive resource on ways students can use AI. "ChatGPT Assignments to use in your classroom today" by Yee et al. (2023) provides more than 60 ways in which students can use AI. It covers topics like Prompt Engineering, Searching, Evaluating, Analyzing, Writing, Generating, and Studying. 

AI Metacognition and Reflection

In a world where Artificial Intelligence (AI) tools are increasingly integrated into all aspects of life, fostering AI metacognition among students becomes paramount. AI metacognition involves students reflecting on and evaluating their interactions with AI technologies to enhance their understanding and application of these tools. This process is crucial not only for optimizing the use of AI in academic tasks but also for developing a critical awareness of how technology influences learning and cognitive processes.

Here are several examples of how students can engage in AI metacognition:

1. Collaborative Dynamics: Reflect on the collaborative dynamic between AI and human intelligence in the learning process. Students should evaluate how AI as a partner contributes to their educational outcomes and identify opportunities for enhancing this cooperation or identifying when AI should not be used.

2. Reflective Comparison: Use AI to compare different drafts of their assignments. This not only helps in recognizing improvements and changes over time but also in understanding how AI suggestions have influenced their revisions and thought processes.

3. Question Generation: Leverage AI to generate questions that probe deeper into their subject matter. This practice can lead to enhanced engagement with the material and a better understanding of the learning content through self-inquiry.

4. Identifying Gaps: Employ AI tools to identify gaps in their knowledge or research. AI can analyze the completeness of their information, suggest areas that require further exploration, and help ensure comprehensive coverage of the topic.

5. Bias Detection and Diversity Enhancement: Use AI to detect potential biases in their work and suggest ways to introduce more diverse perspectives. This reflection is crucial for fostering critical thinking and ensuring the inclusivity of various viewpoints in academic work.

6. Evaluation of Assistance: Reflect on how AI has assisted or hindered their learning process. This includes assessing the accuracy, relevance, and usefulness of the AI-generated content or suggestions.

7. Prompt Evolution: Consider how the prompts they use with AI have evolved or been iterated upon throughout their work. This reflection helps in understanding how their queries and interactions with AI shape the output and its relevance to their objectives.

Engaging students in these metacognitive activities encourages them to become not only users of AI but also thoughtful critics and informed shapers of this technology.

Example AI Assignments

• Wallace - Final Paper Drafting Assignment

• Katzarska-Miller - Group Presentations  

• Wei Lin - Art History Project

• Jackson - Preparing for Debate

• Bandy - Hand on Play and Reflect 

AI Prompting

Talking to AI-powered assistants like ChatGPT is different from what we traditionally think of in terms of searching for information on the internet. Conducting a traditional Boolean search using search engines such as Google is fundamentally different in how AI tools process and respond to queries.

 Traditional Boolean Search

Boolean searches involve using specific operators (AND, OR, NOT) to combine or exclude keywords in a search query. This method is highly structured and requires the user to precisely define what to include or exclude from the search results. For instance, if a student is researching the effects of caffeine on cognitive performance but wants to exclude studies involving adolescents, they might input the following Boolean query into Google: caffeine AND "cognitive performance: NOT adolescents. This search will return documents that mention caffeine and cognitive performance but exclude any documents that mention adolescents.

Talking to AI

Conversely, conversing with an AI involves natural language queries that are more akin to how humans naturally speak. AI uses complex models to understand the intent behind a query and can generate responses based on a wide range of data it has been trained on. This allows for a more intuitive and flexible interaction, where the AI can ask clarifying questions if needed, understand context, and provide answers that are synthesized across various sources of information. For example, a user might ask an AI: Can you explain how caffeine affects cognitive performance in adults? The AI can process this natural language question, understand that the user is interested in the effects of caffeine specifically on adults, and provide a comprehensive summary of the research, possibly drawing from studies, meta-analyses, and its understanding of related topics. 

Key Differences

1. Query Complexity: AI can handle more complex queries expressed in natural language, while Boolean searches require more precise, keyword-based queries structured with logical operators.

2. Interactivity: AI can engage in a dialogue, asking for clarification or providing follow-up information based on user responses, which is beyond the scope of a traditional search engine.

3. Contextual Understanding: AI models generally have a better ability to grasp the context and nuances of a question, whereas Boolean searches strictly match the logical conditions set by the operators. 

Prompting Best Practices

AI prompt engineering, the skill of crafting effective inputs to elicit the most accurate and relevant outputs from AI systems, is an increasingly vital competency for students in the digital age. As AI technologies become more integrated into various sectors of society, the ability to interact efficiently with AI can significantly enhance academic research, problem-solving, and creative projects. Learning to fine-tune prompts helps students achieve better results from AI tools and understand the underlying mechanisms of AI responses. This knowledge promotes a deeper comprehension of AI capabilities and limitations, fostering critical thinking and enabling more informed decisions about when and how to use AI tools. Such skills are essential for navigating the complexities of modern technological environments, making AI prompt engineering a critical addition to educational curricula.

Anthrop\c the creators of Claude and OpenAI the creators of ChatGTP both have resources for how to best communicate with their platforms. However, promoting can be categorized into three main topics: Prompt Specificity, Response Guidelines, and Response Refinement. 

Once these three steps take place students can then spend time validating the AI response with peer review or other relevant literature, further refine the prompt, apply the material to their work, and then cite where AI made its contributions. Reiteration of the AI course policy, citation guidelines, and what constitutes plagiarism should be reiterated when students are applying AI generated content to their original work. 

What if I don't want my students to use AI?

At the end of 2023, about 18% of faculty in higher education either highly restricted or banned the use of AI (Link). The reasons cited were lack of experience, data privacy concerns, ethical considerations, cost/resource constraints, technical challenges, dependence and skill erosion, and lack of evidence for improved learning outcomes. Regardless of the reasons for not using AI, some strategies can help faculty circumvent their students using AI. Below are strategies faculty can use to discourage the use of AI in their classrooms. 

1. Creating an AI Policy - Faculty can set forth guidelines AI use in the syllabus. Ensuring students have this policy on the first day of class can help set the tone for the course and discourage misuse of AI tools. 

2. AI and Plagiarism Discussion - Early in the semester faculty should have an honest discussion about the AI course policy and why they will or will not be using AI, what constitutes AI plagiarism, and what the academic integrity policy is for the institution. Creating a set of community rules with students for the use of AI will help empower them to follow course guidelines.

3. AI Detectors - AI detectors currently exist for text-based generative AI content, such as Turnitin, Drillbit, Google Originality Report, GPTZero, and Grammarly. A comparison/list of detectors can be found here and here, but the most current research states these programs are inconsistent and unreliable (link, link). These detectors fall short when assessing lists, poetry, code, tables, images, and other non-prose content. The best strategy for detecting AI-generated content is to use your expertise, and knowledge of your student's previous work and look for the following in students' written work:

   1. patterns or irregularities in language

   2. check sources and citations

   3. originality, tone, voice, etc. 

   4. factual errors (Hallucinations)

   5. grammar and spelling in AI are usually good, humans make mistakes more often.

For more information, please take a look at the work by Perkins et al (2023) assessing/comparing faculty academic judgment and AI detectors. 

4. Reconsider Assignment/Assessment Strategies - AI struggles with originality, innovation, and creativity in scholarly work (Link). Its strength is in imitation of the content it learns from. Ask students to create individual and personal works that are hard for the AI to replicate. 

   1. Writing-Based Assignments - Use programs that allow for “Track Changes” and “Version History”, Scaffolding Assignments, requiring multiple drafts, Periodic Peer Review, and Reflection Journals (Metacognition) 

   2. Move Up Bloom's Taxonomy - AI is good at: Recalling, Summarizing, Listing, etc. Higher-order tasks make it harder to elicit a nuanced, contextualized, and accurate response from the AI!

   3. Ask for Peer Review Sources - AI does not find reputable peer review resources effectively (Link). By focusing on specific resources or on accurate citations, quotations, and references in work,  AI will struggle and sometimes hallucinate. (Link) 

   4. Focus on How Students Learn - Include assignments that focus on: Hands-on Learning, Active Learning, Field Trips or Experiential Learning, Personal Observation or Reflection, Multimodal Assignments, Group Work, student-led discussion, in-person or video presentation.

   5. Change up Assessment Strategies - Include oral exams, paper exams, video or image-based exams, use time limits, and online proctoring software, and consider ungrading or alternative grading strategies. 

For more information see these resources:

• Dominican University of California (Link)

• University of Chicago (Link)

• University of Illinois Springfield (Link) 

• University of Pennsylvania (Link) 

• Northern Illinois University (Link)

Disclaimer: The content within this compendium was co-created using AI programs ChatGTP and Claude Sonnet. For more information on the co-construction of knowledge using AI, please see this resource by Robertson et al. 2024 and the AI uses in Education Page.