AI Literacy for K-12 Classrooms in the Era of the AI Explosion

Abstract

The article sheds light on how understanding what Artificial Intelligence (AI), including its capabilities, limitations, and influence on our daily lives, is becoming crucial to be integrated into lessons that our students learn every day. It emphasizes that AI literacy needs to be integrated into important literacies students learn in schools such as media literacy, English literacy, numerical literacy, digital literacy, and financial literacy. The article introduces the current work relating to AI literacy done by various scholars and organizations and the pedagogical approaches to promote AI literacy in K-12 classrooms.

Keywords

Cirriculum Planning, Elementary Educaton, Use Modify Create, AI Biases, Secondary Education, AI Literacy

Introduction

To take the best way forward with AI, we’ll have to understand it – understand how to choose the right problems for it to solve, how to anticipate its misunderstandings, and how to prevent it from copying the worst of what it finds in human data. There’s every reason to be optimistic about AI and every reason to be cautious. It all depends on how well we use it. (Shane, 2019, p.235)

When calling customer service, a chat-bot used to be confused if we do not speak standard English. A person with a strong accent had to experience frustration while trying to communicate with a chat-bot. Simultaneous machine interpretation did not make sense most of the time. RoboCup 2017 Symposium used Microsoft’s simultaneous machine interpretation system during its keynote speeches from English to Japanese. After five years, we see astonishing improvement. Now Google Pixel Buds can translate speech in real-time (https://support.google.com/googlepixelbuds/answer/7573100?hl=en). A YouTube video “Two AIs talking about becoming human (GPT-3)” (https://youtu.be/jz78fSnBG0s) shows how AI chatbots can talk to each other to carry out conversations. Although it is not a human-level conversation, they can respond to each other in some meaningful way.

The Artificial Intelligence Index Report 2022 (Zhang et al., 2022) highlighted various trends seen regarding the development of AI in recent years. The private investment in the area of AI has more than doubled in 2021, to around $93.5 billion in total. The total number of AI-related English publications has doubled since 2010 (162,444 in 2010 to 334,497 in 2021). The image clarification systems are becoming more affordable, and the level of their performance is getting better every year. The report points out that language models are growing significantly more capable over years; however, the severity of their biases has intensified. Recommendation algorithms embedded in many the social media including YouTube, Tweets, and Facebook, with a few short clicks, take users from mainstream media to more polarized content provided by hate groups or conspiracy theorists (Shane, 2019). AI algorithms are making decisions for us regarding who gets parole, loans, or is called for an interview by screening resumes. Mostly, their decisions are not impartial. Rather, they can be echoing or amplifying human prejudices which they intend to replace. Furthermore, when AI is misused, intentional or unintentional, or even when it is hacked, there could be larger impacts on our everyday lives (Shane, 2019). AI researchers are aware of issues around AI ethics and research on fairness and transparency (Zhang et al., 2022). Since 2014, the number of publications on AI ethics on the related topics has increased fivefold. In addition, 71% more publications by industry-affiliated researchers contributed to AI ethics-focused conferences. Not only the industry but also governments across the world are involved in the development of AI-related legislation. The analysis of legislative records in 25 countries informed that the number of bills containing Artificial Intelligence passed into law increased from only one in 2016 to 18 in 2021.

What we, the users of AI, need to be aware of is how AI could influence our everyday lives. People selling AI-technology could introduce AI to users as a more capable technology than it actually is or their AI to be impartial while it is measurably biased. Shane points out, “the inner workings of AI algorithms are often so strange and tangled that looking at an AI’s output can be of the only tools we have for discovering what is understood and what it got terribly wrong” (2019, p.4). She explains the AI weirdness as:

• The danger of AI is not that it’s too smart but that it’s not smart enough.
• AI has the approximate brainpower of a worm.
• AI does not really understand the problem you want it to solve.
• But: AI will do exactly what you tell it to. Or at least it will try its best.
• And AI will take the path of least resistance (p.5).

For many of us, especially children, AI could look like magic. It is because much of AI’s decision-making process is hidden in a black box. It is important to inform users that AI is not as smart as it looks, and it will still take many years for the development of Artificial General Intelligence to be accomplished. But then, how can we make sure that AI will not have negative impacts on humans.

Instead of waiting for AI to get smarter, one suggestion is to develop human-AI teams to work collaboratively. The One Hundred Year Study on Artificial Intelligence (AI100) launched in the fall of 2014 aims to provide a long-term investigation of the field of Artificial Intelligence (AI), including its influences on people, their communities, and society (https://ai100.stanford.edu/). They form a Study Panel to evaluate the current state of AI every five years. The 2021 Report, put together by the Study Panel composed of a diverse set of experts in AI highlights the importance of AI approaches that augment human capabilities, through seamless human-machine collaboration and cooperation which enables the decision making aligned with dynamic and complex human values and preferences (Littman et al., 2021). It emphasizes that “an AI system might be better at synthesizing available data and making decisions in well-characterized parts of a problem, while a human may be better at understanding the implications of the data” (p. 9). The report suggests the importance of involving all stakeholders in the development of AI assistance that cooperates and collaborates with humans to produce a “human-AI team that outperforms either alone” (p.9). It is expected that any routine and/or potentially dangerous tasks for humans will be taken over by AI systems. Such applications are possible and effective only when appropriate and inclusive data sets are used while effective integration of algorithms into broader socio-technical systems is achieved.

Integration of AI Literacy in K-12 Education

To establish a society where humans and AI technologies effectively collaborate to enhance each other’s capabilities, not only the people developing AI systems but also everyone who uses such systems needs to have a better understanding of AI and to be involved in decision making regarding where and how to use AI systems. There is a broader understanding among politicians, industries, and educators across fields from computer science, AI, to education that it is crucial to promote AI literacy including the understanding of the science behind AI, its limits, and potential societal impacts in our lives now and in the future. What is specifically urgent is to prepare K-12 students for their future professions, which might not currently exist, and become citizens capable of understanding and utilizing AI technologies in the right way so that they would not benefit some populations over others (Touretzky et al., 2019). This is of particular importance among underrepresented minorities, especially those with low socioeconomic status who have the potential to be left behind due to the lack of resources and opportunities to explore computing in their schools and/or lack of exposure to AI-technologies in their homes and communities.

To promote the integration of AI literacy in K-12 schools, multiple initiatives have already been started. UNESCO’s Unit for Technology and Artificial Intelligence in Education reported three AI curriculum frameworks: AI Literacy: Competencies and Design Considerations, the AI4K12: K-12 AI Guidelines, and the Machine Learning Education Framework (UNESCO’s Unit for Technology and Artificial Intelligence in Education, 2022). It suggests that the three frameworks have the primary purpose of informing the development of AI curricula.

AI Literacy: Competencies and Design Considerations targeting a non-technical audience (Long & Magerko, 2020) present a series of AI competencies and design considerations for AI literacy drawn from existing research. The AI Literacy Competencies Framework contains 17 competencies including:

1. Recognizing AI
2. Understanding intelligence
3. Interdisciplinarity
4. General vs narrow AI
5. AI strengths and weaknesses
6. Imagine future AI
7. Representations
8. Decision-making
9. Machine Learning (ML) steps
10. Human role in AI
11. Data literacy
12. Learning from data
13. Critically interpreting data
14. Action and reaction
15. Sensors
16. Ethics
17. Programmability

AI Literacy Design Considerations include:

1. Explainability
2. Embodied interactions
3. Contextualizing data
4. Promote transparency
5. Unveil gradually
6. Opportunities to program
7. Milestones
8. Critical thinking
9. Identities, values, and backgrounds
10. Support for parents
11. Social interaction
12. Leverage learners’ interests
13. Acknowledge preconceptions
14. New perspectives
15. Low barrier to entry

The Five Big Ideas of AI are AI guidelines for students and teachers developed by the AI4K12 initiative (Touretzky et al., 2019). The Five Big Ideas in AI include (Fig. 1):

1. Perception: Computers perceive the world using sensors,
2. Representation & Reasoning: Agents maintain representations of the world and use them for reasoning,
3. Learning: Computers can learn from data,
4. Natural Interaction: Intelligent agents require many kinds of knowledge to interact naturally with humans, and
5. Societal Impact: AI can impact society in both positive and negative ways.

The Big Ideas 1-4 has a progression chart listing detailed competencies available on their website (https://ai4k12.org/gradeband-progression-charts/).

The Machine Learning Education Framework (Lao, 2020) consists of various knowledge, skills, and attitudes including:

Knowledge

1. General ML knowledge
2. Knowledge of ML methods
3. Bias in ML systems
4. Societal implications of AI

Skills

1. ML problem scoping
2. ML project planning
3. Creating ML artifacts
4. Analysis of ML design interactions and results
5. ML advocacy
6. Independent out-of-class learning

Attitudes

1. Interest
2. Identity and community
3. Self-efficacy
4. Persistence

Lao also suggests that understanding bias and societal implications of AI are the foundation of all skills listed above.

UNESCO’s Unit for Technology and Artificial Intelligence in Education also reported that there are eleven Member States have developed, endorsed, and implemented AI curricula (UNESCO’s Unit for Technology and Artificial Intelligence in Education, 2022). Among the eleven Member States, China, Portugal, Qatar, and United Arab Emirates have AI literacy curricula developed for all educational levels, while ten Member States developed one for high schools. There are four other Member States, including Germany, Jordan, Bulgaria, and Saudi Arabia currently developing AI literacy curricula mainly for middle and high schools.

Its analysis of the curricula revealed the five types of AI integrations: Discrete AI curricula, Embedded AI curricula, Interdisciplinary AI curricula, Multiple-modality AI curricula, and Flexible AI curricula (UNESCO’s Unit for Technology and Artificial Intelligence in Education, 2022). Discrete AI curricula are developed to be an independent subject in a curriculum framework, which has its own time allocations, textbooks, and resources. As of 2022, Only Discrete AI curricula found are China’s Foundations of AI for Grades 10-12 and they are compulsory. Embedded AI curricula are developed to be embedded in other subject area(s). They are mainly embedded in ICT or Computer Science curricula. Some are embedded into language arts, mathematics, science, or engineering. South Korea’s AI curriculum has two AI electives, one in mathematics and the other in technology and home economics. Interdisciplinary AI curricula are developed for cross-subject work through project-based learning involving multiple-subject areas. For example, the UAE’s AI curriculum is integrated into multiple-subject areas including ICT, science, math, language arts, social studies, and moral education. Multiple-modality AI curricula focus on core requirements offered during school time with traditional pedagogical approaches. However, it leverages informal learning opportunities including out-of-school networks and national/international competitions. The IBM-CBSE AI Curriculum for Grade XI & XII offers such a program with a gradual transition from guided to independent learning linking to competitions and industry mentorship. Flexible AI curricula focus on specific integration mechanisms decided by regions, a school district, or individual schools. India’s ATL AI modules curriculum has the flexibility to be offered as embedded, interdisciplinary, or extracurricular programs. Saudi Arabia’s Digital Skills curriculum can be either a discrete or embedded curriculum.

AI Literacy Programs for K-12 Education

Responded to the urgent request to integrate AI literacy in K-12 education, there are a number of open-source curricula and resources developed for middle and high school (AI4ALL, 2021; Clarke, 2019; ISTE, 2021; Payne, 2019). AI4ALL (https://ai-4-all.org) is a nonprofit organization, founded in 2017, dedicated to increasing diversity and inclusion in AI education, research, development, and policy. It has a specific focus on educating underrepresented high school students to become the next generation of AI leaders. Its Open Learning offers various AI lessons for high school students. It provides free curricular materials, teacher resources, and community events available on our online platform. Its curricula align with Next Generation Science Standards (NGSS) engineering standards, the International Society for Technology in Education (ISTE) standards, Common Core ELA/Literacy standards along with CSTA Computer Science standards.

MIT has established RAISE (Responsible AI for Social Empowerment and Education: https://raise.mit.edu/) providing several AI literacy resources including middle school curricula (Lee et al., n.d.; Payne, 2019; Williams, 2021; Williams et al., n.d.). Machine Learning for Kids (https://machinelearningforkids.co.uk/) has a list of AI projects focusing on Machine Learning concepts that teachers can use in their lessons. It also provides teacher log-in where teachers can monitor students’ progress. There are far fewer resources available to elementary schools; however, introducing AI earlier, especially at the elementary grade levels, will inspire young students to learn computer science and acquire computational thinking skills early. Code.org (https://code.org/ai) offers the AI and Machine Learning Module including a video series about AI, AI for Oceans, and AI and ethics, targeting elementary and middle school students. Teachable Machine (https://teachablemachine.withgoogle.com/), Quick Draw! (https://quickdraw.withgoogle.com/) by Google, and GANPaint Studio (https://ganpaint.io/) are some of the many emerging online tools that excite students to learn about AI.

Day of AI (https://www.dayofai.org/) developed and organized by MIT RAISE was held on May 13, 2022, aiming to engage students in a series of hands-on activities designed to introduce Artificial Intelligence (AI) and how it plays a part in their lives. It offered AI activities developed by grade bands (3-5. 6-8, and multiple levels for 9-12). The activities are organized as follows:

1. Learn the Basics: What is AI? What is machine learning? How does it work? How will AI shape my life and that of my community?
2. Identify the Benefits and Risks of AI: What are the social impacts of these technologies? What are important factors to consider in creating AI applications? How can we ensure that AI is used safely and equitably and that people’s privacy is respected?
3. Design and Create: What can I do with AI? How do you train an AI application? What is the future of these technologies?

The total hour for each curriculum is about 4 hours. However, teachers can choose to use some part of the curriculum or divide the curriculum into shorter lessons that teachers can offer in different periods/days.

Pedagogical Approaches and AI Literacy

Many of the AI tools for AI literacy introduce what AI is and help students explore the AI concepts. However, to fully support students’ development of AI literacy, the lessons should include activities that help students to construct new knowledge in meaningful ways on their own by developing AI artifacts or applying the AI knowledge that they have obtained in projects which hold a personal interest for them. Such activities will enhance student learning of AI.

Constructionism theory was developed by Seymour Papert, a student of Piaget, focusing on student learning while constructivism is epistemology – a theory of knowledge. Papert explains:

Constructionism is one of a family of educational philosophies that denies this “obvious truth [teach better].” It does not call in question the value of instruction as such. … Constructionist attitude to teaching is not at all dismissive because it is minimalist – the goal is to teach in such a way as to produce the most learning for the least teaching. (Papert, 1993, p.139)

It focuses on students’ exploration and learning from their own experiences than following the teacher’s instructions. Thus, the lessons should include activities encouraging students’ exploration of learning concepts, allowing students to experience them in their creations.

To support constructionist learning in AI lessons, Lao (2020) recommends employing the Use-Modify-Create (UMC) approach when creating AI artifacts. The UMC approach is often employed in computing education where students first use or engage in exploring a computational tool. They, then, modify it to personalize the creation and finally create their own. It is also important to allow students to follow their interests, providing them student agency – the authorship of their own learning, letting them lead their own learning.  

For example, “An Ethics of Artificial Intelligence Curriculum for Middle School Students” (AI-Ethics) developed by Blakeley H. Payne under the direction of Cynthia Breazeal at MIT Media Lab (Payne, 2019) focuses on promoting the understanding of AI ethics in middle school students. It contains hands-on, mainly unplugged activities to teach various concepts of AI and its ethical influences, both positive and negative, on society. The curriculum consists of eight activities introducing various AI concepts and ethics around AI around social media. The final three projects are built around YouTube as a media platform using the AI-recommender system. “YouTube Scavenger Hunt” encourages students to explore various AI systems used on the YouTube platform. With “YouTube Redesign,” students apply their learning by constructing an ethical matrix around the YouTube recommender algorithm to create a paper prototype of a new version of YouTube that meets the values of the stakeholders identified. The final “YouTube Socratic Seminar” challenges students to discuss which stakeholders were most important or influential in proposed changes to the YouTube Kids app. This lesson provides opportunities for students to experience constructionist learning through various projects and encourages students to gain the authorship of their final products.

AI for CA initiative under the Computer Science for California coalition (CS for CA; https://csforca.org/) developed and offered a professional development workshop gearing up for the Day of AI (https://csforca.org/ai) with AI literacy lessons for K-2, 3-5, 6-8, and 9-12 grade bands. The lessons have a specific focus on students’ AI creation. The lessons are structured around the 4 steps: engage, explore, extend, and explain. Explore is a similar step to modify while extend is the creation of the UMC process following the same principle that students creating AI artifacts should be a critical part of their learning where students gain the ultimate knowledge and skills for their deeper understanding of AI.  

Conclusion

It is not an exaggeration to say that we hear the word AI every day in our lives. No one denies that AI is everywhere. But how many of us can describe what AI is and how it works? Adults including teachers avoid talking about things that we do not fully understand. In the era of the AI explosion, it is imperative that we offer AI literacy to everyone including adults. Furthermore, it is our responsibility to make sure that all students have AI literacy offered in their classrooms either embedded into the core subjects or provided with a multi-disciplinary project-based approach. It is essential for our students to understand how AI technologies work, evaluate their societal impacts, and be empowered to be ethical and responsible AI citizens. While Media literacy education which empowers students to apply critical thinking towards media including social media where various untrustworthy, unreliable, or fake information exists, and to create their own media messages is critical to the health and well-being of our students, AI literacy can provide tools that they can utilize to voice their opinions to create an equitable and inclusive society in the future.

Reference:

AI4ALL. (2021). AI4ALL – Open Learning. https://ai-4-all.org/open-learning/

Clarke, B. (2019). Artificial Intelligence Alternate Curriculum Unit. University of Oregon/Exploring Computer Science. http://www.exploringcs.org/wp-content/uploads/2019/09/AI-Unit-9-16-19.pdf

ISTE. (2021). Artificial Intelligence Explorations and Their Practical Use in Schools. https://www.iste.org/learn/iste-u/artificial-intelligence

Lao, N. (2020). Reorienting machine learning education towards tinkerers and ML-engaged citizens. Massachusetts Institute of Technology.

Lee, I., Ali, S., Zhang, H., DiPaola, D., & Breazeal, C. (n.d.). Developing Middle School Students’ AI Literacy. Proceedings of the 52nd ACM Technical Symposium on Computer Science Education, 191–197.

Littman, M., Ajunwa, I., Berger, G., Boutilier, C., Currie, M., Doshi-Velez, F., Hadfield, G., Horowitz, M. C., Isbell, C., Kitano, H., Levy, K., Lyons, T., Mitchell, M., Shah, J., Sloman, S., Vallor, S., & Walsh, T. (2021). Gathering Strength, Gathering Storms: The One Hundred Year Study on Artificial Intelligence (AI100) 2021 Study Panel Report. Stanford University. http://ai100.stanford.edu/2021-report

Long, D., & Magerko, B. (2020). What is AI literacy? Competencies and design considerations. CHI ’20: Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems, 1–16.

Papert, S. (1993). Mindstorms: Children, Computers, and Powerful Ideas (2nd edition) (2nd ed.). Basic Books.

Payne, B. H. (2019). An Ethics of Artificial Intelligence Curriculum for Middle School Students. MIT Media Lab – AI Education. https://aieducation.mit.edu/aiethics.html (Original source link is broken; visit https://raise.mit.edu/)

Shane, J. (2019). You Look Like a Thing and I Love You: How Artificial Intelligence Works and Why It’s Making the World a Weirder Place. Voracious; Illustrated edition.

Touretzky, D. S., Gardner-McCune, C., Martin, F., & Seehorn, D. (2019). Envisioning AI for K-12: What should every child know about AI? AAAI ’19. AAAI ’19, Palo Alto, CA.

UNESCO’s Unit for Technology and Artificial Intelligence in Education. (2022). K-12 AI Curricula—A mapping of government-endorsed AI curricula. United Nations Educational, Scientific and Cultural Organization.

Williams, R. (2021). How to Train Your Robot: Project-Based AI and Ethics Education for Middle School Classrooms. Proceedings of the 52nd ACM Technical Symposium on Computer Science Education, 1382–1382.

Williams, R., Park, H., Oh, L., & Breazeal, C. (n.d.). Popbots: Designing an artificial intelligence curriculum for early childhood education. Proceedings of the AAAI Conference on Artificial Intelligence, 1, 9729–9736.

Zhang, D., Maslej, N., Brynjolfsson, E., Etchemendy, J., Lyons, T., Manyika, J., Ngo, H., Niebles, J. C., Sellitto, M., Sakhaee, E., Shoham, Y., Clark, J., & Perrault, R. (2022). Artificial Intelligence Index Report 2022. AI Index Steering Committee, Stanford Institute for Human-Centered AI, Stanford University.