Interactive Learning Environments

ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/nile20

Artificial intelligence (AI) literacy education in

secondary schools: a review

Davy Tsz Kit Ng, Jiahong Su, Jac Ka Lok Leung & Samuel Kai Wah Chu

To cite this article: Davy Tsz Kit Ng, Jiahong Su, Jac Ka Lok Leung & Samuel Kai Wah Chu
(08 Sep 2023): Artificial intelligence (AI) literacy education in secondary schools: a review,
Interactive Learning Environments, DOI: 10.1080/10494820.2023.2255228

To link to this article: https://doi.org/10.1080/10494820.2023.2255228

View supplementary material

Published online: 08 Sep 2023.

Submit your article to this journal

Article views: 3069

View related articles

View Crossmark data

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=nile20
INTERACTIVE LEARNING ENVIRONMENTS
https://doi.org/10.1080/10494820.2023.2255228

REVIEW ARTICLE

Artificial intelligence (AI) literacy education in secondary schools:

a review
a a
Davy Tsz Kit Ng , Jiahong Su , Jac Ka Lok Leungb and Samuel Kai Wah Chu a,c

a
Faculty of Education, The University of Hong Kong, Hong Kong; bDivision of Integrative Systems and Design, The
Hong Kong University of Science and Technology, Hong Kong; cSchool of Nursing and Health Studies, Hong Kong
Metropolitan University, Hong Kong

ABSTRACT ARTICLE HISTORY

Artificial intelligence (AI) literacy has emerged to equip students with Received 21 June 2022
digital skills for effective evaluation, communication, collaboration, and Accepted 29 August 2023
ethical use of AI in online, home, and workplace settings. Countries are
KEYWORDS
increasingly developing AI curricula to support students’ technological Artificial intelligence; AI
skills for future studies and careers. However, there is a lack of reviews literacy; review; secondary
summarizing AI education in secondary schools. This study aims to
provide an up-to-date summary of the literature, informing researchers,
policymakers, and educators on fostering students’ AI literacy. A
thematic analysis of 50 AI education studies from 2016 to 2022 was
conducted, evaluating pedagogies, teaching tools, learning contents,
and assessment methods among secondary students. Results were
identified: (1) Collaborative project-based learning, involving
interdisciplinary problem-solving through artifact creation, emerged as
the most common pedagogical approach; (2) Teaching tools were
categorized into hardware, software, intelligent agents, and unplugged
tools; (3) Junior students focused on experiencing AI and basic
concepts, while senior students explored advanced and technical
components; (4) Assessments included knowledge tests, questionnaires,
and qualitative analysis (e.g., videos, documents, presentations); (5)
Students’ learning effects were measured across affective, behavioral,
cognitive, and ethical dimensions. The study identifies suggestions and
challenges for implementing AI education in secondary schools, offering
valuable insights and recommendations for educators and decision-
makers.

1. Introduction
Artificial intelligence (AI) refers to “the science and engineering of making intelligent machines that
have the ability to achieve goals as humans do” (McCarthy, 2007, p. 2). AI technologies have trans-
formed every walk of life (e.g. business, science, art, education) to enhance user experience and
efficiency. People acknowledge the existence of AI applications and devices, but seldom do they
know about the concepts and technology behind them and are aware of their related ethical
issues (Burgsteiner et al., 2016; Ng et al., 2021a, 2021b). Artificial intelligence literacy emerged in
recent years as an important digital skill set in the twenty-first century in order to adapt to the
future (e.g. Druga, Vu, Likhith, & Qiu, 2019; Lee et al., 2021; Long & Magerko, 2020). The term was
first presented by Kandlhofer et al. (2016) to understand the basic knowledge and concepts
behind these AI-driven technologies. It has become an important set of digital competencies that

CONTACT Davy Tsz Kit Ng [email protected]

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10494820.2023.2255228.
© 2023 Informa UK Limited, trading as Taylor & Francis Group
2 D. T. K. NG ET AL.

evaluate, communicate, and collaborate effectively with AI; and use AI as a tool ethically online, at
home, and in the workplace (Long & Magerko, 2020).
Ng et al. (2021a, 2021b) proposed the need for AI literacy education since AI literacy has become an
important digital competency that everyone needs, rather than merely computer scientists. Aligned
with other technological competencies like computational thinking, students were not merely the
end users of the AI technologies, they needed to learn how to solve problems using AI in authentic
settings. As such, teaching and learning strategies are essential to developing young learners’ atti-
tudes, knowledge and skills for their future studies and work. Through hands-on experience, students
can equip themselves with the necessary skills to manipulate AI technologies to help solve problems.
For example, students can use generative AI tools such as ChatGPT to facilitate their writing. Besides,
when editing images, designers can use AI-driven art generators to adjust parameters (e.g. ages,
gender) of people and styles of paintings and photos (Ng, Luo, Chan, & Chu, 2022b). Further,
project-based learning enabled students to build a machine learning model to categorize between
garbage and non-garbage to help clean the ocean using simulation tools (e.g. Teachable Machine,
Code.org AI for Ocean) (Ng et al., 2021a). Successful instruction design and pedagogies can foster stu-
dents’ AI literacy development in terms of attitudes, knowledge, and skills.

1.1. Defining AI literacy education

The terminology surrounding AI in education, AI education and AI literacy is not universally agreed
upon. Scholars debate the appropriate term to use, with some arguing that “AI education” is more
appropriate as it encompasses a wider range of educational activities, while others argue that “AI lit-
eracy education” is more precise and accurate as it emphasizes the development of AI literacy skills
(Eguchi, Okada, & Muto, 2021; Mertala, Fagerlund, & Calderon, 2022). While the two terms may
seem similar, they have distinct meanings (Ng et al., 2023a). AI education is a specialized area in com-
puter science education that involves teaching students about AI knowledge and skills to develop AI-
driven applications (Kandlhofer et al., 2016). However, there is now recognition of the importance of AI
literacy skills for all citizens. Developing AI literacy is increasingly important for people to understand
how AI works, its potential impacts on society, and how to engage with it responsibly (Long &
Magerko, 2020; Ng & Chu, 2021). Since people do not need to learn advanced knowledge such as
algorithms, data structures and programming. As such, some studies tend to use the term “AI literacy
education” which describes the process of teaching students the necessary knowledge and skills to use
AI, develop AI content, and collaborate with others in AI environments through curricula, learning pro-
grams, and pedagogies (Wang & Cheng, 2021). On the other hand, “AI in education” has a different
meaning than the two terms. It refers to the use of AI technology to enhance teaching and learning
in educational settings (Chen et al., 2020; Xie et al., 2021). This can include the use of AI-driven
tools and platforms to personalize learning experiences, assess student performance, provide feedback
to teachers, and automate administrative tasks. To emphasize the importance of developing students’
basic AI knowledge and skills, using the term “AI literacy education” in this review is a more appropri-
ate choice.
Some scholars consider AI literacy as a subset of AI education (Wang & Cheng, 2021), while others
use the terms interchangeably without specifying differences (Chiu & Chai, 2020; Yau et al., 2023).
Additionally, the use of AI technologies in education is not always necessary for AI literacy education.
To illustrate these differences, a comparison between AI education and AI literacy is presented in
Figure 1 (Wang & Cheng, 2021).

1.2. AI literacy education for secondary schools

In school settings, countries have started to design and implement curricula to develop their students’ AI
literacy skills that support their future studies and careers (Ng et al., 2023b; UNESCO, 2022). Discussions
have been raised to understand why all students need to learn AI. For example, AI has generated great
 INTERACTIVE LEARNING ENVIRONMENTS 3

Figure 1. Differences between AI in education, AI education and AI literacy education.

influence to facilitate their learning and living, which is an important digital skill for their everyday life and
future studies. AI causes declines in some occupations and creates great job opportunities in this digital
age (UNESCO, 2022). Therefore, learners need to keep pace with the recent technologies to gain a com-
petitive advantage (Chai et al., 2020; Ng et al., 2021a, 2021b), which raises the need of AI education.
AI literacy is particularly important in secondary school levels, as the focus of this review. Research-
ers stated that secondary students could be more capable of developing a scientific mechanism for
approaching and understanding more complex AI concepts, compared to primary students (Estevez
et al., 2019; Zhang et al., 2022). They could handle more advanced AI knowledge and skills so that
they are able to build AI products to experience different technological roles and collaboratively
make a creation (Ng et al., 2022c). Secondary students are more mature to apply multidisciplinary
knowledge to prepare data, train their models and design authentic products. Students with the
necessary competencies could be more competitive when pursuing AI in their studies and careers
after their graduation (Chiu et al., 2021). With more studies in secondary school settings (e.g. Chiu,
2021; Chiu et al., 2021; Xia et al., 2022), this review summarizes various learning elements in AI literacy
education, including learning content, learning tools, and pedagogies. By gathering effective practices
of AI literacy education in secondary schools identified in this review, we contribute to the develop-
ment of effective elements for instructional and curriculum designs that foster students’ AI literacy.

2. Literature review
This review offers a starting point for researchers and educators in secondary school settings to
understand how to foster students’ AI literacy. This section summarized some scoping reviews in
AI education, and its implementation in K-12 schools.

2.1. Reviews of AI education

In recent years, AI education has become prevalent, and researchers and educators started to discuss
how to teach and learn AI in K-12 education. Teaching AI was not possible in the past because the
4 D. T. K. NG ET AL.

concepts and skills are too advanced and require syntax-based programming skills to develop AI
applications (Wang & Cheng, 2021). Nowadays, with more age-appropriate technologies, secondary
students could visualize AI concepts and learn the working principles behind AI technologies (Kandl-
hofer et al., 2016; Ng et al., 2021a, 2021b; Zhang et al., 2022).
Although AI education is a novel research field, a number of scoping reviews have been identified
in the database that documents the preliminary curriculum design and key concepts of AI education.
Ng et al. (2021a, 2021b) used the Bloom Taxonomy to inform the cognition levels from lower (e.g.
know, understand, use) to higher order thinking (e.g. evaluate, create, collaborate, communicate),
and the technological, pedagogical, and content knowledge (TPACK) model to inform educators
to consider pedagogies, content, and technologies used when designing AI curriculum. Ng et al.
(2022) further identified a set of pedagogies, content, and teaching tools to facilitate teachers to
teach K-16 students AI literacy over the past two decades.
Marques et al. (2020) focused on what types of AI concepts students should learn. The review
identified Touretzky et al. (2019)’s five “big ideas” about AI that every student should know: percep-
tion, representation and reasoning, learning, natural interaction, and societal impact. With techno-
logical advancement, more age-appropriate tools were developed to enable young learners to
learn and experience machine learning through problem analysis, data management, feature engin-
eering, model learning, evaluation, and deployment (Marques et al., 2020).
Sanusi et al. (2021) identified 37 articles to explore the pedagogical approaches to teaching K-12 stu-
dents AI literacy. Problem-based learning, project-based learning, and collaborative learning were pro-
posed to teach high school students machine learning that provides a set of benefits such as
reinforcing students’ higher-order thinking, active learning, cooperation, collaboration, reflection, and
knowledge transfer. Von Wangenheim et al. (2021) presented a ten-year systematic mapping of emer-
ging visual tools to support the teaching of Machine Learning in terms of educational characteristics,
deployment, and how the tools have been developed and evaluated (e.g. Scratch, App Inventor,
mBlock, Teachable Machine, RapidMiner). Sanusi et al. (2021) identified the four main categories of
resources and tools (conversational agents, programming environment, robotic, and unplugged activity)
to document their uses to teach machine learning in K-12 settings from 38 studies.
Tedre et al. (2021) situated the existing 63 documents in the context of computing education to
describe the challenges and 13 important characteristics that teachers/students faced in AI literacy
courses (e.g. focus shift from rules to data, the emphasis on syntax and semantics changes, an age-appro-
priate shift of pedagogical entry points, interaction between body forms and natural language proces-
sing, multidisciplinary applications, societal impact). Further, when talking about AI in recent years, the
learning contents would not merely be rule-based and simple traditional AI. Instead, students could
develop advanced concepts such as machine learning, and natural language (Tedre et al., 2021).
These reviews provide a summary of what educators should teach, what tools should be used,
and what pedagogical approaches should be considered to foster K-12 learners’ AI literacy. These
elements are important for educators and researchers to design appropriate pedagogies, learning
content, and technologies used in their AI curriculum. To specifically understand how AI literacy is
developed for secondary school students, it is meaningful to conduct a review to summarize
these elements (i.e. pedagogies, learning content, technologies, assessment methods, and learning
outcomes) and update the situation for AI literacy education at the secondary level.

2.2. AI education in K-12 schools

AI education for K-12 schools has gained increasing attention among institutions, researchers, and
educators across countries. In the Western world, institutions such as the Association for the
Advancement of Artificial Intelligence (AAAI) and AI4K12.org had developed the “five big” ideas
to start AI learning for K-12 students in the United States in terms of perceptions, representation,
and reasoning, learning, natural interaction and perceptions (Touretzky et al., 2019). Another inter-
national institution called UNESCO announced an education policy “AI in Education: Change at the
 INTERACTIVE LEARNING ENVIRONMENTS 5

Speed of Learning” to describe the current practices of developing and implementing AI curricula in
K-12 schools which include data analytics, personalized learning, and machine learning to explore
the promises of AI in education (UNESCO, 2020). The guideline mentioned the five big ideas in AI
including perception, representation & reasoning, learning, natural interaction, and societal
impact of AI technologies (Touretzky et al., 2019).
In Asian-pacific regions, Su et al. (2022) examined 14 research papers on AI curriculum for K-12 class-
rooms that were taken from 2018 to 2021 by identifying the content knowledge, tools, platforms, activi-
ties, theories and models, assessment methods, and learning outcomes of the selected studies.
Although these institutions have realized the importance of incorporating AI in digital literacy education
for all people, the design of the AI curriculum still needs a rigorous examination to understand the
global standards and how to develop a better pedagogical and instructional design of the AI literacy
curriculum. Of this interest, this review could identify the effective practices for educators, policymakers,
and curriculum designers to follow when developing essential AI concepts, knowledge, and skills.
With more government-endorsed curricula and age-appropriate tools, prior studies have exam-
ined how students scaffold students’ AI literacy. For example, Chiu et al. (2021) created and evalu-
ated a pretertiary AI curriculum to enhance students’ AI awareness, ethics, and impact, knowledge,
and interaction among 335 students in 12-lesson modules. The study found that students could
enhance their AI knowledge and skills, confidence, and intrinsic motivation and have a positive atti-
tude toward AI learning. Ng and Chu (2021) used informal learning to engage students in using
videos, playful tasks and image stylizer via social media. Rodríguez-García et al. (2020) presented
the use of authentic projects to enhance students’ higher cognition levels (e.g. creativity, critical
thinking) using machine learning models to apply their knowledge and propose solutions to their
problems. We can see that more studies have been conducted to measure different aspects of AI
literacy education in terms of pedagogies, tools, and curriculum at the secondary level.
The aforementioned reviews have made important contributions to the field of AI literacy edu-
cation regarding learning content, pedagogy, and teaching tools. However, it is still unclear to edu-
cators how AI has been taught in secondary classrooms in terms of the pedagogies, teaching tools,
learning contents, assessment methods, and learning outcomes. Therefore, in order to understand
the development of AI literacy education, we attempted to fill this gap to explore the field and
suggested future directions regarding how to teach and learn AI in pre-tertiary levels. The following
research questions (RQ) formed the basis of this review:
RQ1: What are the pedagogical strategies used in AI literacy studies?
RQ2: What learning tools have been used in AI literacy studies?
RQ3: What learning contents are appropriate for junior and senior secondary students in AI literacy studies?
RQ4: What assessment methods have been used in AI literacy studies?
RQ5: What learning outcomes have been found in AI literacy studies?

3. Method
3.1. Literature search
This study searched two trusted citation indices databases called the Web of Science and Scopus to
ensure the search covered all the evidence-based SSCI literature. First, we searched the articles pub-
lished in the two databases from 2016 to 2022 according to the phrase (“AI” OR “Artificial Intelli-
gence” OR “Artificial Intelligence literacy” OR “Deep learning” OR “Machine Learning” OR “Neural
network*” OR “Natural language processing” OR “Chatbot”) AND (“K12” OR “K 12” OR “high
school” OR “secondary school” OR “middle school” OR “secondary education” OR “per-tertiary edu-
cation” OR “secondary student” OR “middle school student” OR “per-tertiary student”) AND (“edu-
cation” OR “learning” OR “ teaching” OR “pedagogy” OR “curriculum”) in either the title, the
abstract, main text or keywords were downloaded and reviewed by the researchers. As of 8 Novem-
ber 2022, this gave us a total of 408 articles in the two databases.
6 D. T. K. NG ET AL.

Then, two experienced researchers then determined whether they were suitable for the purpose
of this study. During this search, a set of inclusion and exclusion criteria were set to avoid biases in
the selection of the articles. First, all of the selected studies had to be journal articles, discussion
papers, case studies, or conference papers from the aforementioned databases. Second, the
studies had to be empirical studies in the field of education which was related to AI literacy and
teaching/learning AI concepts (e.g. artificial intelligence, machine learning, deep learning, natural
language, neural networks, chatbots). For example, Sun (2021)’s article was excluded since the
research applied AI technologies in English instruction instead of teaching AI concepts. However,
the editorials and books are excluded due to the lack of peer review. After excluding the irrelevant
studies, a total of 50 articles were identified. An overview of the search protocol is presented in a
PRISMA diagram (Stovold, Beecher, Foxlee, & Noel-Storr, 2014) (see Figure 2). Among the selected
studies, 37 of them were found in both databases, and nine were found in Scopus.

3.2. Data analysis and coding scheme

This article analyzed the background information in terms of countries/regions, educational level, edu-
cational setting, types of articles, and research methods (see Appendix 1). Concerning the research ques-
tions, a theoretical basis was tailored based on Scott’s (2008) curriculum design model that guided the
design of the coding scheme for this review. A set of learning elements were identified including ped-
agogical strategies, learning tools, learning content, assessment methods, and learning outcomes (see
Table 1). Regarding learning outcomes, we also followed the four assessment domains of AI literacy
identified by Ng et al. (2023c), namely (1) affective domains: innate feelings and physiological change
toward AI including intrinsic motivation, self-efficacy, career interest, and confidence; (2) behavioral

Figure 2. PRISMA diagram.

 INTERACTIVE LEARNING ENVIRONMENTS 7

Table 1. Coding scheme.

Themes Descriptions
(1) Years of publications, countries, sample size, research The basic information of the selected studies is discussed,
method, paper type, learner type and data collection including years of publications, countries, and sample size,
method. journal, research method, paper type, learner type and data
collection methods. This aims to understand which countries
and which years have more frequently published AI literacy
studies. Further, the paper type, research method and data
collection methods enable us to understand the strategies and
techniques used in the collection of data to understand how to
teach/learn AI literacy. After that, the sample size and age
distribution of the participants are reviewed
(2) Pedagogical strategies The pedagogical approaches that are often used in the AI literacy
program include collaborative learning, project-based/
problem-based learning, learning with game elements,
scaffolding support, and online learning strategies.
(3) Learning tools This study presents the teaching tools adopted in the AI courses
including: (a) hardware (e.g. CUHKiCar, Micro:Bit, Alpha dog
robot, drones, Raspberry Pi, VR devices), and (b) software (e.g.
Code.org AI Ocean, image stylizer, AI model trainer, AI Scratch,
Machine Learning for kids, Cognimates Colab, WebAPPs,
Google teachable machine, Unity3D, virtual simulation, Smart
Voice, Scratch, Python, IBM Watson Visual Recognition, Snap!,
Cognimates, MIT Ethics Curriculum, ecraft2learn).
(4) Learning content The learning content of the AI concepts and practices are
categorized into lower and upper secondary school levels (e.g.
machine learning, face/image recognition, image stylization,
natural language processing, AI ethics, regression, neural
networks, computer vision, cognition, biomedical informatics,
robotics, information networks, human-robot interactions).
(5) Assessment methods The assessment methods are presented in terms of quantitative
tools (e.g. survey/questionnaire, knowledge tests) and
qualitative methods (e.g. artifact-based assessments,
presentations, interviews, self-reflection).
(6) Learning outcomes Affective, behavioral, cognitive and ethical learning dimensions
are discussed (e.g. interest, motivation, engagement,
satisfaction, behavioral intention, collaboration, knowledge
and skills acquisition).

domains: students’ learning behaviors, operational performance and external actions that demonstrate
collaboration, and behavioral intention and engagement; (3) cognitive learning: students’ lower to
higher cognition skills in Bloom’s taxonomy; and (4) ethical learning: students’ mindsets to address
the ethical issues and concerns about AI, and societal impacts. Figure 3 summarizes the two models
into the theoretical framework for this review.
Inductive thematic analysis was adopted and categorized content with similar meanings into
themes regarding the set of learning elements. To enhance reliability, an inter-rater agreement
was calculated. The first and second authors had more than 80% inter-rater agreement for the
research questions with Cohen’s kappa coefficient (0.85) which demonstrated good inter-rater
reliability (O’connor & Joffe, 2020).

4. Results and discussion

4.1. Data charting and collation
This section presents the educational background of the 46 selected studies identified from the data-
bases (see Table 2).
Years of publication. In the first number of years, the publications of AI literacy papers at secondary
levels are few (9 articles) in the first three years from 2016 to 2019. Forty-one articles were found
between 2020 and November 2022. It is identified that there has been a sharp increase since
2020 (see Figure 4). This trend is consistent with a prior review conducted by Ng et al. (2022a)
8 D. T. K. NG ET AL.

Figure 3. Theoretical framework for this review (Ng et al., 2023c; Scott, 2008).

that 2020 is a breakthrough year of AI education. This shows that AI curriculum and course design
has become increasingly important.
Countries/regions. The countries/regions that published two or more AI literacy articles include the
United States (16), Mainland China (8), Spain (6), and Hong Kong China (4). This pattern is consistent
with a former review conducted by Ng et al. (2021a, 2021b) that the four countries/regions led other
places in developing AI curricula. Regarding participants, the mean age of participants was 15.
Educational levels. The AI literacy courses were implemented across various levels. Most articles
focused on the senior level, covering almost two-thirds of the selected studies. Another one-third of
the studies (14) were conducted in a junior setting while two of the studies were teacher education.
Formal/informal learning. A mapping between years of publication and formal/informal settings
was further conducted. It was identified that seven out of the nine articles from 2016 to 2019
were in informal settings (e.g. online learning, after-school programs, out-of-school activities,
summer program, preliminary experience) (e.g. Gong et al., 2018; Reyes et al., 2020). Nearly half of
the studies (46.0%) from 2020 to 2022 tried to incorporate AI into the computer science curriculum
in regular lessons. Overall, half of the studies were implemented in formal (20) and informal settings
(24). A possible reason for this is that AI literacy is an emerging field, and most researchers tend to
conduct preliminary studies to explore their learning effectiveness in an informal setting.
Research methods. Among the 50 selected studies, there are ten conceptual articles and 40 empiri-
cal studies. Regarding the research method, most of the empirical studies adopted quantitative (17)
and qualitative approaches (11) to examine students’ AI understandings in terms of attitude, motiv-
ation, interest, behavioral intention, and learning performance. Eleven articles adopted a mixed-
method approach to collect data via multiple data sources including knowledge and skills tests,
surveys/questionnaires, observations, field notes, and interviews. Since AI literacy emerged recent
few years, nine conceptual articles were found to adopt an exploratory approach to discuss how
to teach and learn AI in terms of characteristics, tools, and preliminary conceptual frameworks.

4.2. RQ1: what are the pedagogical strategies used in AI literacy studies?
This section identifies the major three pedagogical methods and strategies used in the studies:
project/problem-based learning (36), collaborative learning (28), and experiential learning (26) as
shown in Table 3.
 INTERACTIVE LEARNING ENVIRONMENTS 9

Table 2. Frequency (N, %) of the characteristics of the reviewed articles.

Variables Categories N Percent (%)
Year 2016 2 4
2017 1 2
2018 3 6
2019 3 6
2020 11 22
2021 18 36
2022 12 24
Countries USA 16 32
China 9 18
Spain 6 12
Hong Kong China 4 8
Denmark 2 4
Brazil 2 4
Finland 1 2
Germany 1 2
Poland 1 2
Indonesia 1 2
Norway 1 2
Thailand 1 2
Belgium 1 2
UK 1 2
Austria 1 2
India 1 2
Taiwan 1 2
Educational level * Lower secondary (Grade 7-9) 22 44
Higher secondary (Grade 10 or above) 26 52
Teacher education 2 4
Not specified 6 12
Educational setting Formal 20 40
Informal 24 48
Not specified 6 12
Types of articles Journal articles 24 48
Conference papers 26 52
Research methods Qualitative 11 22
Quantitative 17 34
Mixed research 12 24
Conceptual papers 10 20
Remarks: Some studies focus more than one educational levels.

First, project/problem-based learning is the most commonly used pedagogy to foster stu-
dents’ AI literacy. For example, Vachovsky et al. (2016) engaged 24 girls in authentic projects
to learn computer vision, robotics, and natural language processing in a summer camp. In
the survey, 95.8% of students believed that the projects that students built can help society.
Moreover, students believed that the course was interesting (83.3%) and enhanced their confi-
dence in using AI (75%). Rodríguez-García et al. (2020) presented the LearningML projects (a low
floor high ceiling platform to learn machine learning through doing) to bring the fundamentals
of machine learning to students to acquire its knowledge and educate them to become critical-
thinking citizens. Project/problem-based learning is widely applied in AI literacy education
because it provides authentic settings such as building models of athletic moves (Zimmer-
mann-Niefield et al., 2019), meaningful scientific and STEM contexts that could motivate them
to learn AI through a sense of authenticity and real-world applicability (Sakulkueakulsuk
et al., 2018; Wan et al., 2020).
Collaborative learning allows students to learn how to communicate and work with classmates to
gain AI knowledge and manipulate smart devices (e.g. Deng et al., 2021; Gao & Wang, 2019; Gong et al.,
2018). For example, Gong et al. (2018) engaged students to take different roles such as project man-
agers, software designers, hardware designers, and art designers to build smart vehicles in authentic
settings. Another study conducted by Gao and Wang (2019) also asked students to be buyers and
10 D. T. K. NG ET AL.

Figure 4. AI literacy articles published by year.

sellers to find out the problems and shortcomings of intelligent functions in smart home systems.
Kaspersen et al. (2021) asked students to formulate 3–4 students of a group to design ML models
to predict whether a person will vote for a particular political party. Collaborative learning effectively
assisted students to co-construct AI knowledge and provided them with an authentic environment to
master practical skills and communicate with their group mates. Therefore, we can see the potential of
combining the benefits of collaborative learning and project/problem-based learning to effectively
enhance students’ high-order thinking skills such as problem-solving, leadership skills, project man-
agement, and creativity (e.g. Deng et al., 2021; Gong et al., 2018).
The third common method is to engage students via playful and experiential learning experi-
ences to explore what machine learning is via Teachable Machine (Chiu et al., 2021; Tamborg
et al., 2022), Code.org games (Ng & Chu, 2021), intelligent agents, chatbot and Cognimates (Gong
et al., 2018), syntax-based programs (e.g. Python) (Gong et al., 2018; Gunasilan, 2021) and block-
based programs (e.g. Scratch, Snap!) (Estevez et al., 2019; Kahn et al., 2018). These activities
provide students with hands-on experience to explore what AI is, visualize advanced concepts
(Reyes et al., 2020), and build ML models (Sakulkueakulsuk et al., 2018). Such playful tools and

Table 3. Pedagogical approaches for teaching AI.

Percent
Pedagogies Descriptions Sample studies N (%)
Project/problem- Learning approaches to engage students to (Rodríguez-García et al., 2020; 36 72.0
based learning gain knowledge and skills by working to Sakulkueakulsuk et al., 2018; Wan
investigate authentic questions, problems, or et al., 2020)
challenges.
Collaborative This pedagogy allows students to learn how to (Deng et al., 2021; Estevez et al., 2019; 28 56.0
learning communicate and work with classmates to Gong et al., 2018)
gain AI knowledge and manipulate with smart
devices.
Experiential Process of learning by doing through hands-on (Bellas, Guerreiro-Santalla, 26 52.0
learning experiences and reflection; students could Naya, & Duro, 2022; Chiu et al.,
gain better understanding to connect theories 2021; Tamborg et al., 2022)
and knowledge learned in the classroom to
real-world situations.
Remarks: Some of the articles adopt more than one pedagogies.
 INTERACTIVE LEARNING ENVIRONMENTS 11

platforms could offer students interesting and interactive learning experiences to learn AI. Moreover,
these playful activities offer hands-on experience to scaffold their AI understandings. Some studies
further applied minds-on collaborative projects using playful tools to encourage students to make
AI-driven digital and tangible artifacts. By doing so, educators could enable students to reach
higher cognition levels and apply AI skills and knowledge to solve real-world problems for future
learning and career challenges (Chai et al., 2020).
As revealed by the review, these pedagogical strategies could encourage students to develop skills
such as critical thinking, problem-solving, and communication, and help them apply AI concepts and
skills in real-world contexts. By using a combination of these strategies (e.g. collaborative learning,
project-based learning), educators can create an authentic and engaging learning environment that
prepares students to succeed in our AI-powered world. This enables students to apply AI concepts
and skills in real-world contexts, thus helping them to develop a deeper understanding of AI.

4.3. RQ2: what learning tools have been used in AI literacy studies?
With technological advancement, more and more age-appropriate learning artifacts enable students
to visualize the operations of complex concepts which was not possible in the past. Aligned with
Sanusi et al. (2021) and Ng et al. (2021a, 2021b), four main categories of resources and tools were
identified: conversational agents, programming environment, robotics, and unplugged activity.
Table 4 provides a list of examples of using AI learning artifacts ranging to learn AI.
First, hardware and software-focused artifacts engaged students to use physical and digital arti-
facts to make and create AI-driven applications. For example, Chiu et al. (2021) designed a robotic car
called CUHKiCar with built-in AI functions to offer interactive learning experiences for students to do
face-tracking and line-following tasks. It is found that students obtained a significant improvement
in perceived knowledge gain, confidence, motivation, and AI readiness. Another study conducted by
Chai et al. (2020) used a platform in which students could use Alpha dog robots and design algor-
ithms to do recognitions of physical characteristics such as temperature, voice, face, and images. It
enables students to use AI and mathematics knowledge (e.g. calculus, statistics) to build their AI
solutions.
Second, it is identified that most of the hardware and software should be manipulated with built-
in and add-on AI-driven functions that allow students to build machine learning intelligent agents
and machine learning models without needing to program (Chiu et al., 2021; Kaspersen et al.,
2021). This offers learning opportunities for students to lower the learning barrier and access
more advanced concepts which was not possible in the past. In our selected studies, most of
them focused on tool-based learning to provide students with hands-on experiences to learn
basic AI/computer science knowledge, skills, and concepts. The tools enable students to visualize
complex knowledge and encourage them to collect between AI applications and the underlying
knowledge. For example, AI-driven tools such as chatbots (Rodríguez-García et al., 2020), Scratch
and Teachable Machine (Tamborg et al., 2022) encourage students to solve authentic problems
using these tools. Tool-based AI learning enables students to learn functional and critical aspects
of using AI-driven technologies. In addition to learning technical skills and knowledge (e.g. computer
vision, virtual reality, art design) (Gong et al., 2018), it also requires students’ critical thinking skills to
express and apply knowledge, and communicate and collaborate with the tools to solve authentic
problems (Kaspersen et al., 2021).
As suggested by the results, various learning tools, such as hardware-focused artifacts, software-
focused artifacts, intelligent agents, and unplugged activities, are used in AI literacy studies to
enhance students’ learning experiences and develop their AI literacy. These tools provide hands-
on learning opportunities, enable students to experiment with AI applications that interact with
humans, and offer a fun and interactive way to learn AI concepts and applications (Gunasilan,
2021; Norouzi et al., 2020). Using appropriate learning tools can offer cognitive and affective
12 D. T. K. NG ET AL.

Table 4. Learning artifacts.

Definition Learning artifacts examples Sample studies
Hardware- Use physical artifacts to learn AI such CUHKiCar, educational robot, drones, VR Chai et al. (2020); Chiu
focused as robotics, sensors and Arduino devices Alpha dog robot, Raspberry Pi et al. (2021); Tamborg
artifacts devices. Raspbian, four wheel drive chassis, et al. (2022)
Micro:Bits
Software- Use digital artifacts to learn AI such as Prolog, image stylizer, Colab (Python Chiu et al. (2021); Norouzi
focused block/syntax-based programming programming), Jupyter Notebooks, et al. (2020); Gunasilan
artifacts and simulation. Blockly, WebAPPs, cognitive services, (2021)
Google Collaborator, Scratch, Snap!,
MIT STEP Lab, game design
Intelligent Use intelligent agents such as expert Scratch, Google’s Teachable Machine, Rodríguez-García et al.
agents systems, machine learning trainers, LearningML platform, ecraft2learn, (2020); Sabuncuoglu
chatbots to build their custom Machine Learning for Kids, (2020); Wan et al.
machine learning models without Cognimates Code.org AI Ocean, Face (2020)
coding. AI Scratch, Machine Learning for kids,
AI model trainer, ecraft2learn,
SmileyCluster, chatbots
Unplugged Use learning activities to learn AI Lectures, career talk, textbook, case Gunasilan (2021); Deng
without a computer such as lecture, study, webinar, role-playing, et al. (2021); Gong et al.
case study, role-playing and storytelling, debating (2018)
storytelling.

supports that make complex AI concepts more understandable, and enhance students’ motivation
and confidence (Rodríguez-García et al., 2020).

4.4. RQ3: what learning contents are appropriate for junior and senior secondary
students in AI literacy studies?
This section presents a summary of learning contents that secondary students need to master in
junior (22) and senior levels (26) (see Table 5). To design developmentally appropriate instruction,
educators should understand the cognitive development of each stage when learning AI. In the
junior AI curriculum, prior studies documented that junior secondary students should focus on pre-
liminary and simple AI concepts such as machine learning, natural language processing, and Turing
tests (Chiu et al., 2021; Fernández-Martínez et al., 2021; Ng & Chu, 2021). Educators should design
experiential learning activities for students to have a hands-on experience to taste and use the
related AI applications, discuss their benefits, challenges, ethical concerns and shortcomings of
these tools (Hsu, Abelson, Lao, & Chen, 2021 Sabuncuoglu, 2020). In this way, students needed to
apply these knowledge and skills to solve problems using well-defined hardware, software and intel-
ligent agents.
Senior secondary students could further reach higher cognition levels (Vachovsky et al., 2016)
to develop the technical algorithms and components (e.g. Fisher’s exact test, inductive reasoning,
nearest neighbor algorithm, correlation, graph search). They could taste more advanced concepts
such as computational game theory, agent-based modeling, probabilistic reasoning and graph
theories (Estevez et al., 2019; Reyes et al., 2020). Likewise, after knowing and understanding
the AI knowledge, concerts and skills process, students from both levels could create AI-driven
solutions and models. Moreover, it is noticed that both the junior and senior levels required stu-
dents to consider humanistic, societal and ethical implications of technologies (Kaspersen et al.,
2021). Further, all levels need to learn the important competencies (e.g. critical
thinking, communication, collaboration, creativity) in the twenty-first-century skills that are essen-
tial for students to succeed in school settings and their future workplace (Fernández-Martínez
et al., 2021).
In light of these findings, designing appropriate learning content for junior and senior secondary
students in AI literacy studies is critical for preparing them for developing the necessary competen-
cies to solve problems. For example, junior students should experience AI and gain simple AI
 INTERACTIVE LEARNING ENVIRONMENTS 13

Table 5. Learning contents in junior and secondary school levels.

Levels Learning contents samples Sample studies
Junior Experiencing AI: Using AI applications, benefits, and disadvantages of Chiu et al. (2021); Fernández-Martínez
level using AI, machine learning, face recognition, image stylizer, machine et al. (2021); Ng and Chu (2021)
generation of creative content, experimentation of using AI
technologies;
Simple AI concepts: History/introduction/recent development of AI
and its subareas, differences between humans and machines, natural
language processing, Scratch activities on machine learning and image
recognition, machine learning;
Societal impacts and AI ethics: AI ethics, societal impacts of AI,
machine reasoning, problem-solving
Senior Complex AI topics: Natural language processing, computer vision, Kaspersen et al. (2021); Kahn et al.
level cognition, biomedical informatics, robotics, information networks, (2018); Kong et al. (2022)
human-robot interactions, computational sustainability;
Technical components (for AI education): Fisher’s exact test, inductive
reasoning, nearest neighbor algorithm, correlation, graph search
algorithms, computational game theory, optimization, agent-based
modeling, probabilistic reasoning;
Learning components (for AI literacy): Understanding how ML works,
the process behind creating ML models, and the ability to reflect on its
personal and societal implications.

understanding, whereas senior students can taste more complex AI concepts or even some technical
components for their future study. The learning content should be aligned with the key dimensions
of AI literacy, including knowledge and understanding of AI concepts and applications, ability to
analyze and evaluate AI systems and their outputs, ethical considerations related to AI technology,
and practical skills related to the development and use of AI technology (Ng et al., 2021a; Su et al.,
2023). With this model, educators should choose appropriate levels of knowledge, concepts, and
skills to take care of students’ learning needs and development.

4.5. RQ4: what assessment methods have been used in AI literacy studies?
Among the non-conceptual papers, researchers adopted quantitative (17), qualitative assessments
(11), and mixed methods (12) to examine how students foster their AI literacy-related skills (see
Table 6).

Table 6. Assessment constructs and tools to evaluate students’ AI learning.

Research
methods Constructs and tools (Ng et al., 2021b) Examples Sample studies
Quantitative Use knowledge tests to assess students’ AI What is the importance of similarity when Wan et al. (2020).
(14) cognitive gain and abilities clustering a dataset?
Use perceived questionnaire to assess The content of this AI class is relevant to my Chai et al. (2020);
students’ learning outcomes, including: interests. Chiu et al. (2021).
perceived ability, confidence in using AI, In this AI class, I prefer the materials that
intelligence, truthfulness, perceived really challenge me so that I can learn
understanding, subjective norms, AI new things.
anxiety, perceived usefulness of AI, AI for
social good, attitude toward using AI,
confidence in learning AI, behavioral
intention, AI optimism, relevance, AI
awareness, career adaptability skills.
Qualitative Use videos, documents, pictures, From students feedback, students reported Gunasilan (2021);
(27) presentations, students interactions with that they were excited and happy to try Lee et al. (2021);
AI agents and projects to examine an activity they’ve never tried before in a Druga (2019).
students’ AI affective, behavioral and classroom and most had an underlying
cognitive learning. sense of sportsmanship to prove each
other’s views right/wrong
14 D. T. K. NG ET AL.

Quantitative methods. To understand secondary students’ AI literacy development, surveys and

questionnaires (20) were designed to examine students’ attitudes (e.g. I find using AI technology
to be enjoyable) and perceived abilities (e.g. I have general knowledge about how AI is used
today) (Chiu et al., 2021; Zimmermann-Niefield et al., 2019). Knowledge tests (10) were designed
to assess students’ knowledge acquisition (e.g. Why do you think the large amounts of data
might matter?) (Fernández-Martínez et al., 2021; Norouzi et al., 2020). Among the studies, Chiu
et al. (2021) designed a questionnaire to understand students’ intrinsic motivation, AI readiness, per-
ceived abilities, and confidence. A 41-item questionnaire was developed by Chai et al. (2020) in
which students were examined in terms of their AI literacy development, subject norms, anxiety, per-
ceived usefulness of AI, AI for social good, attitude and confidence in using AI, and behavioral inten-
tion. These questions enabled teachers to examine students’ perceptions of AI literacy education.
Sakulkueakulsuk et al. (2018) examined students’ accuracy rate of the machine learning models to
categorize the quality of mangoes throughout different phases, and other aspects like futuristic
thinking, engagement, interactivity, and interdisciplinary thinking were measured using surveys.
Qualitative methods. Qualitative data was collected by capturing students’ work, field notes,
project presentations, and interviews to examine students’ satisfaction, motivations, and cognition
levels. For example, Lee et al. (2021) revealed students’ daily reflections and interviews about their
workshop experiences to express their excitement when working on this activity. Through inter-
views, students were also able to internalize their learning experience, connect to ethical impli-
cations of technology design (e.g. How accurate is your model? What is the potential for bias?),
challenges of learning AI, and their future job prospects (e.g. What will your future career be?).
Sintov et al. (2016) designed a final project of a role-playing board game to reinforce students’
decision-making. Druga (2019) recorded students’ interaction with AI agents through field obser-
vations and adopted a three-attribute AI perception questionnaire to evaluate how 102 children
(7-12 years old) interacted with and perceived their AI agents in their lessons. These three attributes
measure whether the agents are smarter, truthful, and understand them (e.g. “What do you think of
Google Voice, an AI-enabled agent?”). Children replied that the most fun features were playing
beatbox and music, taking pictures, and playing games. Gunasilan (2021) collected feedback after
debating activities to reflect, and refine the instruction design in an evaluation session in terms of
competitiveness, enjoyment, teamwork, self-reflection, and peer assessment.
Based on the results, a common method of assessing students’ AI literacy development is through
the use of surveys and questionnaires. These surveys typically ask them to indicate their level of atti-
tudes, knowledge, and understanding of AI concepts, and their ability to analyze and evaluate AI
systems (e.g. Chai et al., 2020; Chiu et al., 2021). Another approach to assessing AI literacy is
through the use of performance-based assessments with qualitative analysis. These assessments
can allow researchers to gather rich, detailed information about students’ experiences with AI
and their understanding of its concepts and applications. These methods can triangulate with quan-
titative findings to gain a more comprehensive understanding of students’ AI literacy levels and
identify areas for improvement.
4.6. RQ5: what learning outcomes have been found in AI literacy studies?
Studies adopted various interventions to foster students’ AI knowledge, attitude, and skills. To
examine students’ learning outcomes throughout the learning program, studies measured a set
of learning outcomes and analyzed the effects of learning content, tools, and pedagogies on AI lit-
eracy. The selected studies have reported many meaningful approaches for facilitating students’
learning in AI literacy education. For example, the embedded ethical approach (Grosz et al., 2019),
the project-based AI + ethical learning approach (Williams et al., 2022), the inquiry-based pedagogi-
cal approach of using digital story writing (Ng et al., 2022), and the philosophical reflective model
(Demissie, 2015).
As suggested by Ng et al. (2023), learning outcomes can be categorized into affective, behavioral,
cognitive, and ethical dimensions. Affective learning discusses how students enhance their innate
 INTERACTIVE LEARNING ENVIRONMENTS 15

feelings and physiological change toward AI (Rogaten et al., 2019). For example, a number of studies
examined students’ interest, attitude, confidence, and motivation in an AI education course (Xia
et al., 2022; Zhang et al., 2022). Xia et al. (2022) proposed a questionnaire to measure students’ AI
confidence, readiness, attitude, anxiety, and intrinsic motivation to improve the effectiveness of
the AI curricula for students. Behavioral learning demonstrates students’ behavior-related aspects
such as active learning, behavioral intention, collaboration, and engagement throughout the learn-
ing experience (Jung & Lee, 2018). Norouzi et al. (2020) examined how students enhance their col-
laboration and communication skills via a pre- and post-survey whereas
Priya, Bhadra, Chimalakonda, and Venigalla (2022) evaluated students’ behavioral intention and
interaction throughout their AI learning experience.
Cognitive learning can be understood as a range from lower to higher-order thinking skills in
Bloom’s taxonomy (Ng et al., 2021a, 2021b). Studies are consistent with the AI-focused Bloom’s tax-
onomy to examine learners’ perceived AI knowledge, with a focus on three categories: (1) “knowing
and understanding” refers to the ability to know, explain, exemplify, and understand concepts (Ng
et al., 2021; Zhang et al., 2022); (2) “using and applying” means the competencies to apply prior
knowledge, and develop AI-driven programs and algorithms (Williams et al., 2022; Xia et al.,
2022); and (3) “evaluating and creating” enables students to design and implement AI-driven sol-
utions for problem-solving (Ng, Luo, Chan, & Chu, 2022). Ethical learning can be considered as a criti-
cal knowledge and skill that fosters students to become an ethical and responsible learners
(Kong, Cheung, & Zhang, 2022; Ng et al., 2023). Prior studies integrated ethical awareness (e.g. fair-
ness, accountability, transparency) in AI lessons which aimed to nurture students to become respon-
sible digital citizens (Hsu, Abelson, & Van Brummelen, 2022; Norouzi et al., 2020). Figure 5 developed
a framework that summarized the aforementioned dimensions of learning outcomes for measuring
AI literacy. By measuring these dimensions, researchers can gain a better understanding of the effec-
tiveness of AI education and literacy programs, as well as identify areas for improvement (Chiu et al.,
2021; Xia et al., 2022; Zhang et al., 2022). This proposed framework provides a useful guide for future
research in the field of AI education to measure students’ learning effects and design assessment
tools.

4.7. Contribution of this study

As AI is influencing our everyday life, AI education has been proposed in recent years to help stu-
dents foster their knowledge, skills, and attitude about AI and ethical understanding (Chai et al.,
2020; Kong et al., 2022; Long & Magerko, 2020). When embracing new technologies in educational
settings, educators need always to consider instructional design to facilitate students’ knowledge
and skills acquisition. Although there is much research investigating the effectiveness of pedagogies,
learning content, and teaching tools via various assessment methods, few studies summarize these
elements specific to secondary education. Since teaching AI is new to most secondary school edu-
cators, they may not know the important instructional elements behind it. This study contributes to
summarizing the most commonly used pedagogies and teaching tools to foster students’ different
learning content, knowledge, and skills.
According to Ng et al. (2021b)’s TPACK model, a set of pedagogical (e.g. collaborative learning),
and technological knowledge was listed for educators to equip with so as to facilitate students’ AI
literacy development. This review provides empirical evidence supporting pedagogies and teaching
tools that educators need to know and learn. Regarding tools, teachers could learn how to use hard-
ware and robotics kits (e.g. drones, Raspberry Pi sensors, and Micro:Bits) and software (e.g. Scratch,
Google’s Teachable Machine) as teaching tools to develop students’ AI concepts. They also need to
adopt various intelligent agents (e.g. expert systems, machine learning trainers, chatbots) to build
their custom machine learning models. These tools can effectively complement existing pedagogical
methods (e.g. playful experience, collaborative learning) for educators to design meaningful curri-
cula to develop students’ related knowledge, skills, and attitude to facilitate their study and work
16 D. T. K. NG ET AL.

Figure 5. Affective, behavioral, cognitive and ethical learning outcomes.

(Ng et al., 2022). Educators need to think about how to incorporate these elements according to the
curriculum content, learning goals, students’ characteristics, and learning environment (Ng et al.,
2023a).
To summarize these elements into a framework, Ng et al. (2023a) proposed the various teachers’
inputs for instructional design in terms of learning and teaching (pedagogies), digital resources

Figure 6. Instructional design for AI literacy education.

 INTERACTIVE LEARNING ENVIRONMENTS 17

(technologies), and assessment methods to empower students’ learning. With these instructional
inputs, teachers can consider the age-appropriate learning content from lower to higher cognition
levels, and AI ethics to design activities and learning materials for their students. After all, teachers
can evaluate students’ learning competencies in terms of affective, behavioral, cognitive, and ethical
dimensions. Figure 6 displays the instructional design for AI literacy education.

5. Conclusion
Due to the fourth industrial revolution, artificial intelligence (AI) has become popular and is widely
used across industries thanks to increased data volumes, advanced algorithms, and improvements in
computing power and storage (Reed & Dongarra, 2015). After several years of educational
implementation, AI curricula have been popular to enable students to use AI knowledge and
related technologies to facilitate their learning and build their creations.
This is the first review summarizing the existing evidence of AI education in secondary school set-
tings in terms of research backgrounds, methodological approaches, pedagogical strategies used in
the AI courses, learning tools that are used in the AI courses, learning content, assessment methods,
and learning outcomes. We noticed that our findings aligned with the current reviews (e.g. Marques
et al., 2020; Ng et al., 2021a, 2021b) that we could adapt the Bloom Taxonomy to understand the
cognition gains of AI knowledge, concepts, skills, as well as the TPACK model to understand the
instructional design of choosing appropriate technologies/tools, pedagogies and learning contents
to teach students AI. Furthermore, some studies mention that AI literacy should not merely focus on
cognitive knowledge and skills. Instead, some affective, socio-emotional skills, life and career skills,
and learning skills (e.g. communication, collaboration, critical thinking, creativity) are also important
for students when developing their AI literacy (Ng et al., 2022, 2023). In other words, AI literacy
should be a wider skill set that brings up digital citizens in today’s global community (Ng et al.,
2023). On top of knowing AI concepts and using AI applications ethically, students need to be
able to share and create their thoughts, ideas, and solutions to solve future challenges and
enhance their competitiveness. This review contributes to giving a summary of the up-to-date litera-
ture to inform researchers, policymakers, and educators how to best foster students’ AI literacy at a
per-tertiary level.
Several limitations were identified. First, since most of the articles (21) were conference papers
and half of the articles (15) used qualitative research methods. It was understandable that AI literacy
is still an emerging issue and most research was exploratory in nature. We foresee that future
research design will shift to be more empirical and use rigorous research methods (e.g. quasi-exper-
iment, design-based research) using interventions and control groups. A richer data analysis (e.g. t-
test, ANOVA, factor analysis, regression, structural equation modeling) should be applied. Second,
there are few questionnaires to measure secondary students’ AI literacy and none of them was vali-
dated. Future research could develop AI literacy assessments, especially surveys, and questionnaires,
and examine the reliability and validity of the scales. Finally, to advance this research field, priority
needs to be placed on designing theoretical and pedagogical frameworks to guide policymakers,
educators, and instructional designers with age-appropriate pedagogies, learning artifacts, and
assessment methods.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Mr. Davy Ng is the IT Panel Convener at local secondary school in Hong Kong and a PhD candidate in the Faculty of
Education, the University of Hong Kong. He holds a MEd Educational Psychology, BS Computer Science and
18 D. T. K. NG ET AL.

Postgraduate in IT Education from the Chinese University of Hong Kong. His research interests lie in the areas of AI lit-
eracy, STEM Education and technology-enhanced pedagogic innovation. It is informed by recent research on defining
AI literacy, online learning, and STEM education. He has also served as a reviewer for Artificial Intelligence Review, Com-
puters and Education: Artificial Intelligence, Educational Technology & Society, Information and Learning Sciences and
Journal of Information Technology Education: Research
Ms. Maggie Jiahong Su is currently a PhD candidate in the Faculty of Education at University of Hong Kong. Her research
interests focus on technology education, AI, and STEM in early childhood education. Her publication covers the areas of
artificial intelligence and STEM. She has also served as a reviewer for Computers and Education: Artificial Intelligence, Edu-
cation and Information Technologies, Early Child and Development Care, and Early Childhood Education Journal.
Dr. Jac Ka Lok Leung is currently a Lecturer at the Hong Kong University of Science and Technology. Dr Leung was
awarded Bachelor (Hons) of Mechanical Engineering and Master of Environmental Engineering from The Hong Kong
Polytechnic University and The Hong Kong University of Science and Technology, respectively. In 2021, he was
awarded Doctor of Education at The University of Hong Kong.at the Hong Kong University of Science and Technology.
Dr Leung was awarded Bachelor (Hons) of Mechanical Engineering and Master of Environmental Engineering from The
Hong Kong Polytechnic University and The Hong Kong University of Science and Technology, respectively. In 2021, he
was awarded Doctor of Education at The University of Hong Kong. In Jac’s early career, he worked in the environmental
industry with focus on odor and wastewater contamination. Since 2012, Jac has started working in the higher education
sector, he led and taught in experiential learning programmes and courses for students with diverse backgrounds. He is
dedicated to promoting a “learning-by-doing” approach situated in makerspaces. He and his team were nominated for
the UGC Teaching Award in 2020.
Dr. Chu is an Associate Professor at the Faculty of Education, The University of Hong Kong (HKU). He has obtained 2
PhDs in Education – one focusing on e-Learning from University College London, Institute of Education (2017) and
another one focusing on Information and Library Science from HKU (2005). His areas of expertise include AI literacy,
gamified learning, 21st Century Skills, and Social Media in Education. He has been involved in over 70 research projects
with a total funding of US$ 9,391,342. He has published more than 400 articles and books, with over 100 of them
appearing in international academic journals. Furthermore, Dr. Chu is the Co-Founder and Co-Editor for the journal
Information and Learning Sciences.

ORCID
Davy Tsz Kit Ng http://orcid.org/0000-0002-2380-7814
Jiahong Su http://orcid.org/0000-0002-9681-7677
Samuel Kai Wah Chu http://orcid.org/0000-0003-1557-2776

References
Bellas, F., Guerreiro-Santalla, S., Naya, M., & Duro, R. J. (2022). AI curriculum for European high schools: An embedded
intelligence approach. International Journal of Artificial Intelligence in Education, 1–28.
Burgsteiner, H., Kandlhofer, M., & Steinbauer, G. (2016, March). Irobot: Teaching the basics of artificial intelligence in high
schools. In Proceedings of the AAAI conference on artificial intelligence (Vol. 30(1)). https://doi.org/10.1609/aaai.v30i1.
9864
Chai, C. S., Lin, P. Y., Jong, M. S. Y., Dai, Y., Chiu, T. K., & Huang, B. (2020, August). Factors influencing students’ behavioral
intention to continue artificial intelligence learning. In 2020 international symposium on educational technology (ISET)
(pp. 147–150). IEEE.
Chen, X., Xie, H., Zou, D., & Hwang, G. J. (2020b). Application and theory gaps during the rise of artificial intelligence in
education. Computers and Education: Artificial Intelligence, 1, 100002. https://doi.org/10.1016/j.caeai.2020.100002
Chiu, T. K. (2021). A holistic approach to the design of artificial intelligence (AI) education for K-12 schools. TechTrends, 65
(5), 796–807. https://doi.org/10.1007/s11528-021-00637-1
Chiu, T. K., & Chai, C. S. (2020). Sustainable curriculum planning for artificial intelligence education: A self-determination
theory perspective. Sustainability, 12(14), 5568.
Chiu, T. K., Meng, H., Chai, C. S., King, I., Wong, S., & Yam, Y. (2021). Creation and evaluation of a pretertiary artificial intelli-
gence (AI) curriculum. IEEE Transactions on Education.
Demissie, F. (2015). Promoting student teachers’ reflective thinking through a philosophical community of enquiry
approach. Australian Journal of Teacher Education, 40(12), 1–15. http://doi.org/10.14221/ajte.2015v40n12.1.
Deng, W., Huang, X., Liu, Q., & Wang, Z. (2021, December). Curriculum design of artificial intelligence in middle school-
taking posture recognition as an example. In 2021 tenth international conference of educational innovation through
technology (EITT) (pp. 310–315). IEEE.
 INTERACTIVE LEARNING ENVIRONMENTS 19

Druga, S., Vu, S. T., Likhith, E., & Qiu, T. (2019). Inclusive AI literacy for kids around the world. Proceedings of FabLearn,
2019, 104–111.
Eguchi, A., Okada, H., & Muto, Y. (2021). Contextualizing AI education for K-12 students to enhance their learning of AI
literacy through culturally responsive approaches. KI-Künstliche Intelligenz, 35(2), 153–161.
Estevez, J., Garate, G., & Graña, M. (2019). Gentle introduction to artificial intelligence for high-school students using
scratch. IEEE Access, 7, 179027–179036. https://doi.org/10.1109/ACCESS.2019.2956136
Fernández-Martínez, C., Hernán-Losada, I., & Fernández, A. (2021). Early introduction of AI in Spanish middle schools. A
motivational study. KI-Künstliche Intelligenz, 35(2), 163–170. https://doi.org/10.1007/s13218-021-00735-5
Gao, J., & Wang, L. (2019, August). Reverse thinking teaching discussion in high school information technology under
new curriculum standards. In 2019 14th international conference on computer science & education (ICCSE) (pp. 222–
226). IEEE.
Gong, X., Wu, Y., Ye, Z., & Liu, X. (2018, June). Artificial intelligence course design: iSTREAM-based visual cognitive smart
vehicles. In 2018 IEEE intelligent vehicles symposium (IV) (pp. 1731–1735). IEEE.
Grosz, B. J., Grant, D. G., Vredenburgh, K., Behrends, J., Hu, L., Simmons, A., & Waldo, J. (2019). Embedded EthiCS: inte-
grating ethics across CS education. Communications of the ACM, 62(8), 54–61.
Gunasilan, U. (2021). Debate as a learning activity for teaching programming: A case in the subject of machine learning.
Higher Education, Skills and Work-Based Learning.
Hsu, T. C., Abelson, H., Lao, N., & Chen, S. C. (2021). Is it possible for young students to learn the AI-STEAM application
with experiential learning?. Sustainability, 13(19), 11114.
Hsu, T. C., Abelson, H., & Van Brummelen, J. (2022). The effects on secondary school students of applying experiential
learning to the Conversational AI Learning Curriculum. International Review of Research in Open and Distributed
Learning, 23(1), 82–103.
Jung, Y., & Lee, J. (2018). Learning engagement and persistence in massive open online courses (MOOCS). Computers
and Education, 122, 9–22.
Kahn, K. M., Megasari, R., Piantari, E., & Junaeti, E. (2018). AI programming by children using snap! block programming in
a developing country. https://ecraft2learn.github.io/ai/publications/EC-TEL_2018_source-files_48%20kk%20edits%
20changes%20accepted.pdf
Kandlhofer, M., Steinbauer, G., Hirschmugl-Gaisch, S., & Huber, P. (2016, October). Artificial intelligence and computer
science in education: From kindergarten to university. In 2016 IEEE frontiers in education conference (FIE) (pp. 1–9).
IEEE.
Kaspersen, M. H., Bilstrup, K. E. K., Van Mechelen, M., Hjorth, A., Bouvin, N. O., & Petersen, M. G. (2021, June).
Votestratesml: A high school learning tool for exploring machine learning and its societal implications. In
FabLearn Europe/MakeEd 2021-An International Conference on Computing, Design and Making in Education, 1–10.
Kong, S. C., Cheung, W. M. Y., & Zhang, G. (2022). Evaluating artificial intelligence literacy courses for fostering concep-
tual learning, literacy and empowerment in university students: Refocusing to conceptual building. Computers in
Human Behavior Reports, 7, 100223.
Lee, I., Ali, S., Zhang, H., DiPaola, D., & Breazeal, C. (2021, March). Developing middle school students’ AI literacy. In
Proceedings of the 52nd ACM technical symposium on computer science education (pp. 191–197). https://doi.org/10.
1145/3408877.3432513
Long, D., & Magerko, B. (2020, April). What is AI literacy? Competencies and design considerations. In Proceedings of the
2020 CHI conference on human factors in computing systems (pp. 1–16).
Marques, L. S., Gresse von Wangenheim, C., & Hauck, J. C. (2020). Teaching machine learning in school: A systematic
mapping of the state of the art. Informatics in Education, 19(2), 283–321. https://doi.org/10.15388/infedu.2020.14
McCarthy, J. (2007). From here to human-level AI. Artificial Intelligence, 171(18), 1174–1182. https://doi.org/10.1016/j.
artint.2007.10.009
Mertala, P., Fagerlund, J., & Calderon, O. (2022). Finnish 5th and 6th grade students’ pre-instructional conceptions of
artificial intelligence (AI) and their implications for AI literacy education. Computers and Education: Artificial
Intelligence, 3, 100095.
Ng, D. T. K., & Chu, S. K. W. (2021). Motivating students to learn AI through social networking sites: A case study in Hong
Kong. Online Learning, 25(1), 195–208.
Ng, D. T. K., Lee, M., Tan, R. J. Y., Hu, X., Downie, J. S., & Chu, S. K. W. (2022a). A review of AI teaching and learning from
2000 to 2020. Education and Information Technologies, 1–57.
Ng, D. T. K., Leung, J. K. L., Su, M. J., Yim, I. H. Y., Qiao, M. S., & Chu, S. K. W. (2023a). AI literacy in K-16 classrooms. Springer
International Publishing AG. https://link.springer.com/book/10.1007/978-3-031-18880-0
Ng, D. T. K., Luo, W., Chan, H. M. Y., & Chu, S. K. W. (2022b). Using digital story writing as a pedagogy to develop AI literacy
among primary students. Computers and Education: Artificial Intelligence, 3, 100054.
Ng, D. T. K., Su, J., & Chu, S. K. W. (2023b). Fostering Secondary School Students’ AI Literacy through Making AI-Driven
Recycling Bins. Education and Information Technologies.
Ng, D. T. K., Wu, W., Leung, J. K. L., & Chu, S. K. W. (2023c). Artificial Intelligence (AI) literacy questionnaire with confi-
rmatory factor analysis. 23rd IEEE International Conference on Advanced Learning Technologies, US.
20 D. T. K. NG ET AL.

Ng, D. T. K., Leung, J. K. L., Chu, K. W. S., & Qiao, M. S. (2021a). Ai literacy: Definition, teaching, evaluation and ethical
issues. Proceedings of the Association for Information Science and Technology, 58(1), 504–509. https://doi.org/10.
1002/pra2.487
Ng, D. T. K., Leung, J. K. L., Chu, S. K. W., & Qiao, M. S. (2021b). Conceptualizing AI literacy: An exploratory review.
Computers and Education: Artificial Intelligence, 2, 100041. https://doi.org/10.1016/j.caeai.2021.100041
Norouzi, N., Chaturvedi, S., & Rutledge, M. (2020, April). Lessons learned from teaching machine learning and natural
language processing to high school students. In Proceedings of the AAAI conference on artificial intelligence (Vol.
34(09), pp. 13397–13403). https://doi.org/10.1609/aaai.v34i09.7063
O’connor, C., & Joffe, H. (2020). Intercoder reliability in qualitative research: Debates and practical guidelines.
International Journal of Qualitative Methods, 19, 1609406919899220.
Priya, S., Bhadra, S., Chimalakonda, S., & Venigalla, A. S. M. (2022). ML-Quest: a game for introducing machine learning
concepts to K-12 students. Interactive Learning Environments, 1–16.
Reed, D. A., & Dongarra, J. (2015). Exascale computing and big data. Communications of the ACM, 58(7), 56–68. https://
doi.org/10.1145/2699414
Reyes, A. A., Elkin, C., Niyaz, Q., Yang, X., Paheding, S., & Devabhaktuni, V. K. (2020, August). A preliminary work on visu-
alization-based education tool for high school machine learning education. In 2020 IEEE integrated STEM education
conference (ISEC) (pp. 1–5). IEEE.
Rodríguez-García, J. D., Moreno-León, J., Román-González, M., & Robles, G. (2020, October). Introducing artificial intelli-
gence fundamentals with LearningML: Artificial intelligence made easy. In Eighth international conference on techno-
logical ecosystems for enhancing multiculturality (pp. 18–20). https://doi.org/10.1145/3434780.3436705
Rogaten, J., Rienties, B., Sharpe, R., Cross, S., Whitelock, D., Lygo-Baker, S., & Littlejohn, A. (2019). Reviewing affective,
behavioural and cognitive learning gains in higher education. Assessment and Evaluation in Higher Education, 44
(3), 321–337.
Sabuncuoglu, A. (2020). Designing one year curriculum to teach artificial intelligence for middle school. In Proceedings of
the 2020 ACM conference on innovation and technology in computer science education (pp. 96–102.
Sakulkueakulsuk, B., Witoon, S., Ngarmkajornwiwat, P., Pataranutaporn, P., Surareungchai, W., Pataranutaporn, P., &
Subsoontorn, P. (2018, December). Kids making AI: Integrating machine learning, gamification, and social context
in STEM education. In 2018 IEEE international conference on teaching, assessment, and learning for engineering
(TALE) (pp. 1005–1010). IEEE.
Sanusi, I. T., Oyelere, S. S., Agbo, F. J., & Suhonen, J. (2021, October). Survey of resources for introducing machine learn-
ing in K-12 context. In 2021 IEEE frontiers in education conference (FIE) (pp. 1–9). IEEE.
Scott, D. (2008). Critical essays on major curriculum theorists. Routledge.
Sintov, N., Kar, D., Nguyen, T., Fang, F., Hoffman, K., Lyet, A., & Tambe, M. (2016). From the lab to the classroom and
beyond: Extending a game-based research platform for teaching AI to diverse audiences. Proceedings of the AAAI
Conference on Artificial Intelligence, 30, 1.
Stovold, E., Beecher, D., Foxlee, R., & Noel-Storr, A. (2014). Study flow diagrams in Cochrane systematic review updates:
An adapted PRISMA flow diagram. Systematic Reviews, 3, 1–5.
Su, J., Guo, K., Chen, X., & Chu, S. K. W. (2023). Teaching artificial intelligence in K–12 classrooms: A scoping review.
Interactive Learning Environments, 1–20.
Su, J., Zhong, Y., & Ng, D. T. K. (2022). A meta-review of literature on educational approaches for teaching AI at the K-12
levels in the Asia-pacific region. Computers and Education: Artificial Intelligence, 3, 100065.
Sun, X. (2021). 5G joint artificial intelligence technology in the innovation and reform of university English education.
Wireless Communications and Mobile Computing, 2021, 1–10.
Tamborg, A. L., Elicer, R., & Spikol, D. (2022). Programming and computational thinking in mathematics education: An
integration towards AI awareness. KI-Künstliche Intelligenz, 36(1), 73–81.
Tedre, M., Toivonen, T., Kaihila, J., Vartiainen, H., Valtonen, T., Jormanainen, I., & Pears, A. (2021). Teaching machine learn-
ing in k-12 computing education: Potential and pitfalls. arXiv preprint arXiv:2106.11034.
Touretzky, D., Gardner-McCune, C., Martin, F., & Seehorn, D. (2019, July). Envisioning AI for K-12: What should every child
know about AI?. In Proceedings of the AAAI conference on artificial intelligence (Vol. 33(01), pp. 9795–9799). https://doi.
org/10.1609/aaai.v33i01.33019795
UNESCO. (2020). AI in education: Change at the speed of learning. https://iite.unesco.org/wp-content/uploads/2020/11/
Steven_Duggan_AI-in-Education_2020.pdf
UNESCO. (2022). K-12 AI curricula: A mapping of government-endorsed AI curricula. https://unesdoc.unesco.org/ark:/
48223/pf0000380602
Vachovsky, M. E., Wu, G., Chaturapruek, S., Russakovsky, O., Sommer, R., & Fei-Fei, L. (2016, February). Toward more
gender diversity in CS through an artificial intelligence summer program for high school girls. In Proceedings of
the 47th ACM technical symposium on computing science education (pp. 303–308). https://doi.org/10.1145/
2839509.2844620
Von Wangenheim, C., Hauck, J. C., Pacheco, F. S., & Bertonceli Bueno, M. F. (2021). Visual tools for teaching machine
learning in K-12: A ten-year systematic mapping. Education and Information Technologies, 26(5), 5733–5778.
https://doi.org/10.1007/s10639-021-10570-8
 INTERACTIVE LEARNING ENVIRONMENTS 21

Wan, X., Zhou, X., Ye, Z., Mortensen, C. K., & Bai, Z. (2020, June). Smileycluster: Supporting accessible machine learning in
K-12 scientific discovery. In Proceedings of the interaction design and children conference (pp. 23–35). https://doi.org/
10.1145/3392063.3394440
Wang, T., & Cheng, E. C. K. (2021). An investigation of barriers to Hong Kong K-12 schools incorporating artificial intelli-
gence in education. Computers and Education: Artificial Intelligence, 2, 100031. https://doi.org/10.1016/j.caeai.2021.
100031
Williams, R., Ali, S., Devasia, N., Dipaola, D., Hong, J., Kaputsos, S. P., Jodan, B., & Breazeal, C. (2022). AI ethics curricula for
middle school youth: Lessons learned from three project-based curricula. International Journal of Artificial Intelligence
in Education, 1–59.
Xia, Q., Chiu, T. K., Lee, M., Sanusi, I. T., Dai, Y., & Chai, C. S. (2022). A self-determination theory (SDT) design approach for
inclusive and diverse artificial intelligence (AI) education. Computers and Education, 189, 104582.
Xie, H., Hwang, G. J., & Wong, T. L. (2021). Editorial note: From conventional AI to modern AI in education: Reexamining
AI and analytic techniques for teaching and learning. Journal of Educational Technology & Society, 24(3), 85–88.
Yau, K. W., Chai, C. S., Chiu, T. K., Meng, H., King, I., & Yam, Y. (2023). A phenomenographic approach on teacher con-
ceptions of teaching artificial intelligence (AI) in K-12 schools. Education and Information Technologies, 28(1), 1041–
1064.
Zhang, H., Lee, I., Ali, S., DiPaola, D., Cheng, Y., & Breazeal, C. (2022). Integrating ethics and career futures with technical
learning to promote AI literacy for middle school students: An exploratory study. International Journal of Artificial
Intelligence in Education, 33, 290–324.
Zimmermann-Niefield, A., Turner, M., Murphy, B., Kane, S. K., & Shapiro, R. B. (2019, June). Youth learning machine learn-
ing through building models of athletic moves. In Proceedings of the 18th ACM international conference on interaction
design and children (pp. 121–132). https://doi.org/10.1145/3311927.3323139