UDL + AI = OMG: How to use Artificial Intelligence to wow your physics students with Universal Design for Learning

Universal Design for Learning (UDL) is a framework that guides the design of learning experiences to proactively meet the needs of each and every learner. UDL is based on three principles that provide multiple means of engagement, representation, and action and expression. In this blogpost, I will share how artificial intelligence (AI) can be used to support and enhance UDL in the physics classroom, as well as some of the ethical and social implications of using AI in education. I will also discuss how AI can be used for developing creative assessment. The examples I will use are from the perspective of a senior physics teacher who is teaching NCEA Physics in New Zealand.

AI and multiple means of engagement

Engagement is the affective aspect of learning that involves motivation, interest, and emotional regulation. AI can help provide multiple means of engagement by offering personalized and adaptive learning experiences, gamified and interactive elements, and feedback and scaffolding.

• Personalized and adaptive learning: AI can use data and algorithms to tailor the learning content, pace, and difficulty to each learner’s needs, preferences, and goals. For example, Knewton is an AI-powered learning platform that can recommend physics topics and activities based on the learner’s prior knowledge, performance, and interests, and adjust the level of challenge and support accordingly.

• Gamified and interactive elements: AI can also make learning more fun and engaging by incorporating gamified and interactive elements, such as points, badges, leaderboards, quests, and avatars. For example, Physics Playground is an AI-powered physics game that can simulate realistic and immersive scenarios, such as launching a rocket or building a bridge, and reward the learner for applying physics concepts and skills.

• Feedback and scaffolding: AI can also provide timely and personalized feedback and scaffolding to the learner, such as hints, explanations, and prompts. For example, Socratic is an AI-powered physics tutor that can monitor the learner’s progress and performance and provide feedback and guidance on how to solve physics problems, or how to improve their physics understanding.

AI and multiple means of representation

Representation is the cognitive aspect of learning that involves perception, comprehension, and memory. AI can help provide multiple means of representation by offering diverse and multimodal learning resources, natural language processing, and visualization and modeling.

• Diverse and multimodal learning resources: AI can help curate and generate diverse and multimodal learning resources, such as texts, images, videos, audio, and animations, that cater to different preferences. For example, an AI-powered physics textbook called Google Bard can provide different formats and modalities of presenting physics concepts and phenomena, such as text, diagrams, graphs, equations, videos, simulations, and narrations. Google Bard can also generate physics poems and stories to make learning more enjoyable and memorable.

• Natural language processing: AI can also use natural language processing (NLP) to analyze, understand, and generate natural language, such as speech and text. For example, an AI-powered physics assistant called ChatGPT can use NLP to answer physics questions, explain physics concepts, or generate physics summaries, using natural and conversational language. ChatGPT can also engage in physics-related dialogues with the learner and provide feedback and suggestions.

• Visualization and modeling: AI can also use visualization and modeling to help learners perceive, comprehend, and manipulate physics concepts and phenomena. For example, an AI-powered physics simulator called Physion can use visualization and modeling to create dynamic and interactive representations of physics systems, such as forces, motion, energy, and waves, and allow learners to explore and experiment with different variables and parameters. Physion can also be used to create any kind of mechanism and simulate it in real-time.

AI and multiple means of action and expression

Action and expression are the behavioral aspect of learning that involves communication, production, and self-regulation. AI can help provide multiple means of action and expression by offering flexible and creative ways of demonstrating learning, automated and formative assessment, and metacognitive and reflective tools.

• Flexible and creative ways of demonstrating learning: AI can help learners express their physics understanding and skills in flexible and creative ways, such as writing, speaking, drawing, coding, or designing. For example, Makeblock is an AI-powered physics project that can challenge learners to use their physics knowledge and creativity to design and build a physics-related product, such as a robot, a solar panel, or a roller coaster, and present their work using different media and formats.

• Automated and formative assessment: AI can also help assess learners’ physics understanding and skills in an automated and formative way, such as providing instant and adaptive feedback, scoring and grading, and generating reports and recommendations. For example, Quizizz is an AI-powered physics quiz that can use automated and formative assessment to measure learners’ physics knowledge and skills, provide feedback and hints, score and grade their responses, and generate reports and recommendations for further learning and improvement.

• Metacognitive and reflective tools: AI can also help learners develop their metacognitive and reflective skills, such as planning, monitoring, and evaluating their physics learning. For example, Nurture is an AI-powered assessment tool designed to use metacognitive and reflective tools to help students plan their physics learning goals and strategies, monitor their physics learning progress and performance, and evaluate their physics learning outcomes and achievements. 

Ethical and social implications of using AI in the physics classroom.

While AI can offer many benefits and opportunities for enhancing UDL in the physics classroom, it also poses some ethical and social challenges and risks that need to be addressed and mitigated. Some of these challenges and risks include:

• Privacy and data protection: AI relies on collecting and analyzing large amounts of data from learners, such as their personal information, learning behaviors, and performance. This raises concerns about how the data is collected, stored, shared, and used, and whether the learners’ privacy and data rights are respected and protected. For example, who owns and controls the data? How is the data secured and anonymized? How is the data used and for what purposes? How can learners' access, modify, or delete their data?

• Bias and fairness: AI can also be biased and unfair, as it can reflect and amplify the biases and inequalities that exist in the data, algorithms, and systems. This can result in discrimination and harm to learners, especially those from marginalized and underrepresented groups. For example, how are the data and algorithms representative and inclusive of the diversity of learners? How are the outcomes and decisions of AI transparent and accountable? How are the impacts and consequences of AI monitored and evaluated?

• Agency and autonomy: AI can also affect the agency and autonomy of learners and teachers, as it can influence and interfere with their choices, actions, and interactions. This can result in a loss of control and responsibility, as well as a lack of critical thinking and creativity. For example, how are the roles and relationships of learners and teachers changed by AI? How are the learners and teachers involved and empowered in the design and use of AI? How are the learners and teachers encouraged and supported to think critically and creatively with and about AI?

AI and creative assessment in physics education

Creative assessment is a form of assessment that requires learners to demonstrate their learning in novel and original ways, such as creating, designing, or inventing something. Creative assessment can foster and measure learners’ higher-order thinking skills, such as analysis, synthesis, and evaluation, as well as their creativity, innovation, and problem-solving skills. AI can be used to support and enhance creative assessment in physics education, by providing tools and resources for generating, evaluating, and improving creative products and processes. Some examples of how AI can be used for creative assessment in physics education are:

• Generating creative physics problems: One way to do this is to use a large language model (LLM) like ChatGPT that has been trained on a large corpus of physics texts and questions. An LLM can generate novel physics problems by using various prompting techniques, such as zero-shot, in-context learning, and confirmatory checking. For example, an LLM can be given a prompt like “A ball is thrown at an angle of 30 degrees with an initial speed of 20 m/s. What is the maximum height reached by the ball?” and then asked to generate a similar but different problem by changing one or more parameters or variables. The LLM can also provide feedback and hints to the students by comparing their answers with the correct ones. This way, AI can enhance physics learning by providing students with diverse and engaging physics problems that stimulate their curiosity and creativity.

• Evaluating creative physics solutions: AI can be used in a physics classroom to evaluate creative physics solutions by providing feedback, guidance, and alternative approaches to students’ work. For example, Virtual Reality Simulations in Physics Education, which allows students to learn physics concepts such as wave propagation, ray optics, relative velocity, electric machines, etc. in a high school or college level. The application uses a three-dimensional interface and interactive animations to create realistic and engaging physics scenarios for students to explore and manipulate. The application also provides feedback and assessment to help students monitor their learning outcomes. Additionally, the application can generate reports and analytics that help teachers and students evaluate their progress and performance. By using AI as a tool-to-learn-with, students can enhance their physics understanding and develop their creativity and problem-solving skills.

• Improving creative physics thinking: There are many AI tools that can be used to create or evaluate physics solutions, but one example is Smodin, a powerful AI assistant that can simplify solving challenging physics equations and provide step-by-step solutions. 

Reflection and conclusion

In this blogpost, I have shared how AI can enhance UDL in the physics classroom, by providing multiple means of engagement, representation, and action and expression. I have also discussed some of the ethical and social implications of using AI in education, such as privacy and data protection, bias and fairness, and agency and autonomy. Finally, I have discussed how AI can be used for developing creative assessment in physics education, by providing tools and resources for generating, evaluating, and improving creative products and processes.

I hope this blogpost has given you some insights and ideas on how AI can be used to support and enhance your physics teaching and learning. I also hope this blogpost has prompted you to reflect on your own teaching practice, and how you can integrate AI in your physics classroom in a meaningful and ethical way. I invite you to share your thoughts and experiences with me and other educators in the comments section below.

If you are interested in exploring these ideas further, consider registering for the following course from Ed3DAO and Idaho State University: Learning Stream