This post is written by guest contributor Ian G. Williams.

AI literacy is having a moment. Since the release of ChatGPT 3.5 in November 2022, widespread use and adoption of artificial intelligence systems – specifically large language models, generative (pre-trained transformer) AI systems (genAI) – have skyrocketed around the world. Often, this adoption is integrated and layered into existing infrastructures and enterprise systems: Microsoft Copilot suddenly appearing in your Outlook inbox, a notice from Google that Gemini is now part of your storage plan, a note at the top of Adobe Acrobat that says it’s easy to summarize a long article and save time, Brightspace offering to generate quiz questions. New models and versions of AI systems seemingly appear every day: AI agents taking the technophile world by storm, and “AI theater” becoming a real-life instance of Pokémon arguing things out on a social media site. The pace of development and deployment continues. The material and social infrastructures to support AI – data centers, high power computing, fiber optic cables, coding schools and boot camps, AI forward policies – receive massive investments.

With this explosion of genAI growth, a potpourri of neologisms has emerged – terms like “prompt engineering,” “vibe coding,” and “model switching.” These denote ideas of a skillful use of the technologies; a capacity to instrumentalize and coax them towards a desired set of outcomes – a prototype, a website, an image, an essay, a travel itinerary. These terms imply a mastery and synergy with these new interfaces, and a capacity to integrate them into everyday life. These ideas broadly fit within the concept of “digital literacy,” a term first introduced in 1997 by Paul Gilster, which supplanted its predecessor term, “computer literacy.” Digital literacy was intended to be more general and broad, and was often used interchangeably with the concept of new media literacies in the early 2000’s. When the big data revolution kicked off in the 2010s, data literacy was the new thing (Pangrazio and Sefton-Green, 2020). AI literacy’s ascent and emergence are an extension of the lineage, forming around a particular set of technologies, and it is the latest evolution of this ever-changing concept. It’s everywhere because we live in a world saturated with AI systems and agents, discourse, marketing, and ideas about AI. For instance, when riding the subway, take a moment to observe how many ads for AI systems and tools cover the train.

The world is saturated by AI systems, which often get framed as both causes of and solutions to society’s ills. Making meaning of AI is largely framed in terms of literacy; as Luci Pangrazio (2026) argued in a recent essay, AI literacy is invoked in both optimistic and pessimistic responses to AI’s ascent – even as the continued idea of literacy forecloses on other possibilities for relating to, finding meaning in, and resisting the flow of AI. Yet AI literacy is here, shaping not just scholarly inquiry, but also policymaking. Last week, the US Department of Labor released an AI Literacy Framework, formalizing a working definition of what this means, part of a wider commitment of the Federal government to “prioritizing AI literacy and skill development across the workforce and education systems” (Department of Labor, 2026, p. 1-2).

Like technologies before it, the sudden rollout of AI systems into the world through unregulated, direct-to-consumer releases does not equally distribute benefits and harms. AI universalism is limited and ignores the importance of local context, and how AI systems layer into and rearrange existing sociotechnial systems. Examining how AI literacies are enacted and played out in people’s lives can help ground discussions. While many of us in higher education feel pressure to develop and articulate AI literacies to preserve our turf, engage in sustained critique, and adjust to a new reality, we are a small subset of overall users and consumers of digital technologies. It can be easy to lose sight of this in the academy and forget how everyday people of a variety of backgrounds and skillsets experience this AI moment, and this call for AI literacy.

Earlier this week, I attended a talk and then an invited workshop, organized by Data & Society. It focused on the report (404) Job Not Found: The AI Literacy Trap At Work, an ethnographic study by Anuli Akanegbu, PhD, of Atlanta’s digital skills and job landscape in the midst of the “AI literacy” boom. Akanegbu’s report also builds upon Daniel Greene’s work in this area, crystallized in his book The Promise of Access: Technology, Inequality, and the Political Economy of Hope where he argues that in the 1990’s, problems of persistent poverty – particularly in an era of welfare reform and neoliberal market policies – were transformed into problems of technology access: what is now known as the digital divide. Greene describes “the access doctrine” – a belief that access to technology (and requisite skills) can solve poverty and social inequality – in his comparative institutional ethnography of Washington, DC in the mid 2010’s. At a conference last year, Greene explained to me how his project emerged from frustrations working in social services and workforce development, where the mantra “learn to code” was unquestionably repeated as the solution to structural poverty.

Akanegbu’s report builds an important layer of this work, and updates for the current landscape of the mid 20202’s, by taking a close look into Atlanta, a major Southern US city with the highest wealth disparity in the country, the origin of the cotton gin, and a currently booming tech hub that was reinventing itself as “Silicon Peach.”  The report followed the experiences of Black Atlanta residents as they navigated the AI literacy landscape. The event discussed perspectives on the report, including the central argument that AI literacy is a “strategic abstraction” – a deliberately vague concept that is hard to pin down, and always pointing towards an uncertain future, proclaimed in the language of policymakers, businesspeople, and public-private partnerships. This vagueness keeps people, particularly job-seekers and working folk, ever guessing, always having to upskill and position themselves for a changing set of definitions and technologies. Akangegu identitied a digital skills training and AI literacy model that offers an alternative to patchwork, undersupported landscapes often designed with the assumption that participants who are white collar office workers, upskilling on the job, flexibly scheduled, or able to bear the risks of tuition for a precarious and uncertain future. Serena Oduro and Anuli Akakngebu’s companion policy brief identifies strategies for tangible improvements to the current AI literacy landscape.

Our conversation afterwards with a range of stakeholder organizations – policymakers, labor unions, think tanks, educational institutions, workforce development social enterprises – focused on the concept of AI literacy in education, labor policy, and history. It made me think about my own experiences with trying to make sense of this concept, and what the different dimensions of what becoming “AI literate” might entail. Rather than presume an object or universal understanding, I thought it would be helpful in this note to locate and situate my experiences, which are rooted in my activities and relationships at GCDI. Although I am a social worker, my “practice” has been primarily on campus and in university settings for some time now – I work in community with other academics. Much of my funding has come from fellowships and programs associated with GCDI, which exposed me to the world of digital humanities.

I have been thinking and reading about AI literacy, and digital skills, for some time. This genAI moment unfolded during my time at The Graduate Center and my perception has been shaped by institutional context. Sometimes, I am so immersed in how we approach technology, education, and society here that I forget the distinct contours of our approach and our daily practices here – hence why the Data & Society event was so refreshing and thought-provoking. It is hard to articulate concisely what we do at GCDI, but if I had to, I would say our approach grounds humanistic inquiry, tempered optimism, a strong commitment to situated ethics, and a belief in the importance of fostering communities of practice and mutual support might start to sum it up. CUNY has a rich history of both organizing against harmful technological practices in education and building alternative practices and infrastructures (Fabricant & Brier, 2016). There is a real commitment here to seeking structural policy solutions along with adapting education and other institutions to the current times – such as we are seeing in initiatives such as The Critical AI Literacy Institute and the emergent CUNY AI Lab. It gives me some hope regarding how we can develop meaningful and actually useful approaches to what AI literacy can be, and sustains my curiosity about how this AI literacy moment will continue to play out in CUNY, and in society.

—

References

Fabricant, M., & Brier, S. (2016). Austerity blues: Fighting for the soul of public higher education. JHU press.

Gilster, P.  (1997). Digital literacy. New York: Wiley Computer Pub.

Greene, D. (2021). The promise of access: Technology, inequality, and the political economy of hope. mit press.

Pangrazio, L. (2026). The (im)possibility of AI literacy. Learning, Media and Technology, 51(1), 1–7. https://doi.org/10.1080/17439884.2026.2615553

Pangrazio, L., & Sefton-Green, J. (2020). The social utility of ‘data literacy’. Learning, Media and Technology, 45(2), 208-220. https://doi.org/10.1080/17439884.2020.1707223