Interview | Victor R. Lee: Why AI Literacy Is Becoming a Necessity for Every Student

Informally, AI literacy is the knowledge and skills that enable students to decide if, when and how to use AI technologies in their activities with the understanding that it is a tool that has both capabilities and limits. Largely, this is seen as understanding some of the basic mechanisms of how AI works, such as how data is used to train systems to detect and predict patterns; considering if it works for a given task as a complement, support or a less desirable alternative; and knowing the risks and implications with its use that may go beyond the immediate task at hand.

In recent years, a lot of attention was given on how to operate specific AI tools and activities such as ‘prompt engineering’. However, those are really narrow views of AI literacy and most people working in the area would reject those as being defining—in part because the interfaces and specific series are going to change and we are focused on what are enduring ideas that affect how we use technology to affect our world.

Core competencies include understanding how systems of people and technology can be coordinated to work together in ways that are both productive and beneficial and posing problems and tasks that can be iteratively solved and addressed with human ingenuity and modern tools.

The current landscape is complex but the sense is that AI use is already happening even in the absence of clear guidance or policies. Even though there are high levels of use, there is also a lot of wariness and suspicion. In schools, it is a fear of cheating, over-reliance and the devaluing of core human skills. Among the youth, rejection of AI is growing due to larger concerns about the environment and societal inequities. Young people are afraid they won’t develop effectively if they become over-reliant.

Innovative use tends to look at AI as an augmenter of human expressive capabilities rather than as a one-stop tutoring system. Classrooms may pursue ambitious projects, build prototypes, handle public data sets and effectively take on project-based learning. Other innovative uses include critically examining AI—activities where students actively critique it by testing out its capabilities and limitations and thinking about how they need to consider its limitations so they can get the most out of their interactions.

There have been some cases of individual students who use it to discover new topics or ideas that would not have been easily available to them, such as children in rural regions or in small schools where some topics of interest are not offered or available.

If we engage in our learning activities where we do not, on the human side, practice critical thinking, engage in complex problem-solving or exercise and celebrate human creativity, then we would expect those to weaken. With AI, the fear is that those go away because we largely used those to complete the tasks that AI can do well—such as writing an analytical essay or coming up with new solutions or new creative products. Education has been incentivised to value the product more than the process and shifting focus to the process and how we do the work is going to be important. Seeing AI used as a choice rather than as an imperative is also important. There are a huge number of things where humans are far superior to AI and we want to celebrate and encourage those in our schooling because that is going to be a huge part of success and fulfilment as an adult. Practically, that may include some limitations on its use during early ages and selective use in classroom activities and learning. AI may be impressive, but it is just a tool for our activities—not a replacement for our capabilities.

There are topics that are personal and those that are more societal. Personal topics include privacy, good verification practices in a world of deepfakes, security, intellectual property and responsible conduct such as how to be a good digital citizen and prevent activities like cyberbullying or online harassment. Societal topics include understanding how different groups are being affected by AI’s rise in terms of economic opportunity, geopolitical implications of AI [such as military uses or interfering in elections], representation in society, what histories are told and sustainability. Ethics should be integrated into AI rather than as a separate topic—deciding whether to use AI for certain activities involves considering the likely quality of information; what are some tendencies for who is represented because of the training data that were used; whether the resource use for the task merits AI or would be done just as well with a few seconds of individual thinking. It sounds daunting but we do this already in many ways—when we decide what mode of transportation to use for different activities, how to handle waste and recyclable materials, in our purchasing decisions and when we make decisions related to policies in our communities. Schools should model students on how to think through these considerations and help students to reflect on their values and how their actions align with or should change to align with those values.

Much more energy and resources need to go into teacher preparation and training around AI as they are going to be a major point of contact for how students use AI. Teachers’ AI literacy is important and will involve how to teach about AI in contextualised and developmentally appropriate ways. This also involves knowing if and how to use AI effectively for their work. In the US, schools are seeking policy support—there are a lot of issues they are not sure how to approach and want examples and suggestions that work.

Assessment needs to evolve, which will require giving time for teachers and curriculum developers to make those changes. There is an endless game of cat and mouse if we focus only on keeping AI out of assessments and assignments. For some activities, assessment will be done in a low-tech, no-tech way. For others, the assessment is partly about how technology is being used and how students show their own growth given the tools they have around them through portfolios or reflection activities as well as new types of assessable products—students might, for instance, vibe code interactive simulations to demonstrate principles they have learned and show their extensions and personal explorations.

As my work is not in India, I will comment with the caveat that I lack some of the expertise I think would be important here. But, with unequal technology, diversity and teacher shortage, what we should avoid is looking at AI as the solution—as in having AI be a substitute for teachers in a region where there is a shortage or using AI as a reason to consider language diversity. We need to ensure equitable access, which may require infrastructural investments and investing in public spaces—where key ideas about AI and how it can be used—are made available to all. Policymakers should encourage and incentivise industry partners to invest sustainably and impactfully in regions that are not well served currently. Also, “unplugged” methods—teaching about technical topics without using digital technology—can be powerful and still be made available so those should be made available until other infrastructures are established.

A lot depends on decisions being taken now by policymakers and the private sector. The overhead in using AI basically is relatively low and should continue to get lower as companies are incentivised to make their products easy-to-use. For example, learning how to write a computer code can be quite challenging, but with AI, you can use human language to say, “write a program that will take a list of numbers and put them in order from smallest to largest”. Where I see a difference to be concerned about is in the users who know that AI may miss important ideas to consider—like how to work with edge cases or user errors when producing that code—and those who accept whatever AI gives as being right and superior. That is where I see a possible danger. And that extends to dealing with hallucinations, using AI for medical, therapeutic or social support or for learning about the world. Effective use means more than operating AI—it means knowing what it does well and what it doesn’t and what we as humans do to get the most out of it so that we can thrive.

To put a play on words into the conversation, I don’t see AI transforming education so much as education transforming how we think about AI. Right now, AI seems almost magical in what it can do. With education, we can see it for what it is—very impressive computing that could help in certain tasks and be limited for others. If we keep education as the central focus—knowing that AI is one of many things that are changing in our world where we need to be critical, creative and flexible, rather than treating AI as being in the driver’s seat—I think we’ll be in a much better position. Some of that will involve reaffirming what humans are good or better at in our schooling system and fostering a sense of agency where we see AI as providing some powerful tools that can help us to create the world we want to live in rather than a world where it feels like AI is defining a world where we are constantly racing in order to keep up.