Reconsidering literacy in an AI world

America has been obsessed with reading for decades, but the rise of artificial intelligence may make other forms of literacy just as – if not more – important.

• American education has focused on reading for decades, with little effect on student performance.
• The growth of artificial intelligence creates opportunities to access information in different ways that may involve less actual reading.
• Educators need to look ahead to the types of literacy people will need in the future.
• Six literacies that could become important are information access literacy, prompt engineering literacy, literacy for truth, communication literacy, ethical literacy, and the literacy mix.

We should all be embarrassed. We are the education policy makers arguing that the science of reading is better or worse than other approaches. We are the education researchers diligently synthesizing and analyzing the literature about reading and conducting research about teaching reading. We are the school leaders, the school board members, and the parents worrying that third graders can’t read. We are the teachers stuck in the middle of the reading wars, cultural battles, and children who need our care.

We should be embarrassed because we’ve been making practically the same arguments about the teaching of reading for decades. But guess what? Reading proficiency hasn’t changed very much, and children’s reading test scores have hardly improved. According to the National Assessment for Educational Progress (NAEP) for 9-year-olds, after a bump between 1999 and 2004, scores in 2022 were at the same level as 1980. For 17-year-old students, scores rose all of 2 points from 1971 until 2012 (National Center for Education Statistics [NCES], 2021).

The pandemic had some recent impact, but the reality is that after years of intensive focus on reading instruction, we have little to show for it. Indeed, the most recent release of NAEP showed a decline of two points from 2022 for fourth- and eighth-grade students (NCES, 2025).

Our obsession with reading scores should embarrass us because we have arrived at a time when artificial intelligence (AI) is challenging every aspect of humanity. For example, it can easily complete school homework and ace academic exams. It soon will be able to make just about every book ever written available on audio.

We should be embarrassed because today’s technology has made reading largely unnecessary for accessing information and communication. We should be embarrassed because education has so many problems and our children have so many issues to address, yet we are still mired in seeing reading as a key education priority.

The American obsession with reading

Reading is important. But Americans’ obsession with it is unique. No other country is as preoccupied as the U.S. — not even countries where students learn more than one language or whose writing system is much more complex than English. Billions of people around the world learn English as a second language, and billions of people learn to read and write language systems much more complex than English. However, their education systems are not nearly as worried about their children’s basic reading.

Americans’ obsession with reading has as much to do with economics and politics as it does with anything else. First, reading involves a lot of money. The average spending per student in the U.S. was $16,280 for 2020-21 (using constant 2022-23 dollars), with total elementary and secondary school expenditures at $927 billion that year (NCES, 2024). In primary grades, much of the expenditure on the curriculum side is spent on reading and math. When a state or district adopts a reading approach or program, the textbooks, training, assessments, and consultants are costly. The companies that provide these products have a financial interest in being selected. Ironically, one of the best ways to be successful is to discredit competitors and get them out of the market. The big scandal around the federal Reading First program, which was accused of favoritism and conflicts of interest, is a great example of how profits can become a higher priority than actual effectiveness (Mismanagement and Conflicts, 2007).

The reading wars are also political. American education, unlike education in most countries, has always been responsive to politics. And the debate on how to best teach reading has largely been divided by political views. Traditionalists want to make sure all children are taught specific things in a specific way, often at the same pace, leaving children little space for self-development and control. Progressive educators want children to have more space and time to develop reading abilities, to have access to meaningful texts, and to engage in authentic reading as early as possible. Which side dominates doesn’t change the outcomes very much.

Education’s forever war

School systems tend to copy one another in the never-ending race to achieve excellence. States and schools change pedagogies, reading programs, textbooks, and reading consultants every few years to keep up with the latest thinking. None of this necessarily solves the major reading challenges, although it does foster an ongoing war that, amazingly, has spanned more than 50 years (Preston, 2022).

Leaders on both sides claim to have discovered the holy grail or panacea for reading, which of course is not possible. Reading is very personal. Even the latest neuroscience driving much of the current debate, which highlights common processes in reading, offers little in the way of useful translational research for teachers in real classrooms (Shanahan, 2020). And the everyday reality for teachers is that different children have different needs, face diverse challenges, and require varying approaches (Kuhn & Dougherty-Stahl, 2022).

The current battle revolves around what has been labeled the “science of reading,” with high-profile media attention sparking intense interest from politicians, educators, and academics.

In each round of the reading wars, past failures get blamed on the inadequacy of previous approaches, and the “new science” is the panacea for solving literacy concerns. The “bad guys” are usually inadequately trained teachers and the institutions who prepare them for being too dense to implement what the critics believe are the obvious remedies. Lost are all the nuances that affect learning to read — and the fact that the newly banned practices happened to work for some kids. If the prolonged skirmishes have any legacy, it is that learning to read is hard for some students, and solutions that work for everyone are elusive.

Reading in an AI world

Advances in AI raise questions about how central reading will be in the future. And these advances come rapidly. Google’s recent release of its generative AI platform Gemini displayed remarkable advances in interactivity, video, and audio. Its NotebookLM produces human-sounding podcasts in minutes about any uploaded documents. ChatGPT can create audio for just about any topic fed to it with the proper prompts. StormAI creates sophisticated research articles on nearly any subject in short shrift. Kahnmigo offers personalized tutoring and learning assistance to students. Multiple generative AI models can sift through reams of data or videos and offer analyses that would take humans hours and days on their own. The list could go on. What’s next? Nobody knows, except that today’s generative AI is the dumbest it ever will be, and it knows more than any of us already.

But for reading, the 20th- and 21st-century models continue to dominate discourse, even as students are told they are being prepared for jobs that don’t exist yet. Nobody has a clue what skills students will need for these jobs that haven’t been created, yet our teaching and learning are based on the past.

Our thinking is that curiosity, creativity, innovation, critical thinking, and other human-centered abilities will continue to be essential and should dominate what we should teach. Admittedly, none of that is certain either. But we do know that the old models never met everyone’s needs.

What literacies do we need?

For years, literacy scholars have expressed the need for what was called multiple literacies. There have been several iterations of what this means and what types of knowledge and skills make up the necessary literacies. In general, though, multiple literacies are a sort of “toolbox in which literacies are taken up as visual, oral, written, tactile, olfactory, and multimodal digital” (Masny, 2011, p.496).

We envision several literacies that new technology demands. The importance of these new literacies expands what it means to be literate and diminishes the priority of simply decoding and comprehending. It questions defining literacy as scoring well on reading tests.

We suggest six new literacies for consideration.

Information access literacy

We refer to knowing how to best use AI’s emerging source of information and power as information access literacy. Learners in the past had access to information through libraries, newspapers, television, radio, and so on. In recent years, social media outlets, YouTube, podcasts, and others — including AI — offer all sorts of audio and video access points. What kinds of questions, written or oral, are best answered with AI?

Prompt engineering literacy

The ability to use generative AI effectively requires knowledge of how to prompt it for the purposes being pursued. Literacy in prompt engineering means being able to create written or audio statements for a generative AI tool to get the desired kind of response. Prompt engineering expert Lance Eliot (2024) put it this way, “The use of generative AI can altogether succeed or fail based on the prompt you enter.” There are varying approaches, tips, and ideas that can help learners create good prompts. Interestingly, the AI tools themselves can help. For example, Microsoft Copilot offers prompting suggestions.

Literacy for truth

We know that social media is fraught with inaccuracies, and literacy experts already have noted the importance of media literacy as a critical ability to process fake news and discern the truth (Mo Jones-Jang, Mortensen, & Liu, 2021). Digital literacy is another broader term that focuses on processing and analyzing information (Radovanovic et al., 2020). Because AI tools build their responses from information already available, including on social media, they make mistakes, called hallucinations. Strategies such as lateral reading (Wineburg & McGrew, 2017) that work for reviewing human-created material can also work for AI. An approach called chain of verification prompting can be useful. It is a prompting technique used to verify and improve generative AI responses to ensure reliable and accurate responses. With generative AI’s abilities drawn from millions of sources, photos, and videos, ways to validate what is being read, seen, or heard grow. We refer to this as literacy for truth.

Communication literacy

Texting among people is common for communicating, often using standard and non-standard abbreviations, pictures, and emojis, versions of language the unindoctrinated can’t fathom. But it works, especially among younger individuals who often rely on it as a primary means of communication and social interaction. Traditional Reading ability has its place, but being able to read what is written in the texting world is increasingly important.

Ethical literacy

Fears of cheating pervade discussions about using AI. Moreover, questions about when AI use is appropriate and how to cite it are rising among content creators and researchers. Disclosure seems to be a key, as even federal grant agencies have offered suggestions for using AI in research applications. If AI enables students to drive their own education, how might that look, and what ethical responsibilities do teachers, schools, and states have? And finally, as AI-built virtual reality (VR) becomes more pervasive, teacher work will likely change, focusing on the more human components of teaching. What will the parameters be around what is ethical and proper?

The literacy mix

Students will need to be able to comprehend information from multiple sources. Even if one can read the written word and comprehend it, does the same hold for audio, video, and pictures? Each format requires different comprehension skills, and different information-sharing approaches will require literacies for different senses and modalities. We refer to this as the literacy mix, as it captures the understanding that people glean information, and future knowledge will require skill sets likely never considered or emphasized.

The future of literacy

Kids are reading less, and reading skills for all ages seem to be eroding (Crain, 2007; Organization for Economic Cooperation and Development, 2024). Some people even worry that we are becoming a “post-literate” society (O’Connor, 2024). We believe that some level of reading is important, though we don’t know how important it will be in the future or what level of reading ability will be necessary to be truly literate. AI and other technologies are creating an informational and cultural challenge. Shying away from the new reality should be everyone’s biggest fear.

Students at all levels are getting large percentages of their information from non-written venues. Perhaps some mixture of reading with other modern comprehension needs will be key. But the backward-facing conception of literacy driving education today needs to change, and it can’t happen too soon.

Crises always cause changes in how people live and operate. Generative AI, VR, and other technologies put us in information-crisis mode. This requires us to examine past practices and adjust to meet new demands. Will people still read books, whether on paper or online? Likely so, at least in the short term. But the dominance of that mode of information access and communication probably will wane. It is time we faced what confronts us and alter our notions of literacy.

References

Crain, C. (2007, December 16). Twilight of the books. The New Yorker.

Eliot, L. (May 9, 2024). The best prompt engineering techniques for getting the most out of generative AI. Forbes.

Kuhn, M.R. & Dougherty-Stahl, K.A. (2022). Teaching reading: Development and differentiation. Phi Delta Kappan, 103 (8), 25-31.

Masny, D. (2011). Multiple literacies theory: Exploring futures. Policy Futures in Education, 9 (4), 494-504.

Mismanagement and conflicts of interest in the Reading First Program: Hearing before the U.S. House of Representatives Committee on Education and Labor, 110th Cong. (2007) (FirstM).

Mo Jones-Jang, S, Mortensen, T. & Liu, J. (2021). Does media literacy help identification of fake news? Information literacy helps, but other literacies don’t. American Behavioral Scientist, 65 (2), 371-388.

National Center for Education Statistics. (2021). Explore NAEP long-term trends in reading and mathematics. U.S. Department of Education.

National Center for Education Statistics. (2024). Public school expenditures. Condition of Education. U.S. Department of Education.

National Center for Education Statistics. (2025). Explore results for the 2024 NAEP reading assessment. U.S. Department of Education.

O’Conner, S. (2024, December 26). Are we becoming a post-literate society? Financial Times.

Organization for Economic Cooperation and Development. (2024). Do adults have the skills they need to thrive in a changing world? Survey of adult skills 2023. OECD Publishing.

Preston, T. (2022). A look back: A chronicle of Kappan’s coverage of the reading wars. Phi Delta Kappan, 103 (8), 5-7.

Radovanovic, D., Holst, C., Belur, S.B., Srivastava, R., Houngbonon, G.V., Le Quentrec, E., Miliza, J., Winkler, A.S., & Noll, J. (2020). Digital literacy key performance indicators for sustainable development. Social Inclusion, 8 (2), 151-167.

Shanahan, T. (2020). What constitutes a science of reading instruction?Reading Research Quarterly,55 (1), S235-S247.

Wineburg, S. & McGrew, S. (2019). Lateral reading: Reading less and learning more when evaluating digital information. Teachers College Record, 121.

ABOUT THE AUTHORS

Rick Ginsberg

Rick Ginsberg is dean of the School of Education and Human Sciences at the University of Kansas. He is the co-author of Duck and Cover: Confronting and Correcting Dubious Practices in Education (Teachers College Press, 2023).

Yong Zhao

Yong Zhao is Foundation Distinguished Professor at the University of Kansas School of Education, in Lawrence. His most recent book is Duck and Cover: Confronting and Correcting Dubious Practices in Education (Teachers College Press, 2023).