Will Anyone Still Need Me? A Small Consultant’s Stake in 2035

Christopher Ross

School of Education and Technology, Royal Roads University

LRNT 523: Foundations of Learning and Technologies

Dr. Elizabeth Childs

October 25, 2026

Will Anyone Still Need Me? A Small Consultant’s Stake in 2035

I run a one-person WordPress shop. Most days, my actual job is teaching small business owners enough about their own websites and their own AI tools that they stop needing me for every small thing. That is the whole model: I am paid, in part, to make myself less necessary. So when I was asked to imagine 2035, I did not reach for a classroom. I reached for my own inbox.

A Tuesday in 2035: a client messages me with a question about her website traffic. Before she finishes typing, her AI assistant has already answered it, correctly, in the time it would have taken her to open my email. She still messages me, this time just to say thanks for teaching her enough to trust the answer. That version of 2035 I can live with. The one I worry about is the one where she stops messaging me at all, because the assistant was there first and good enough, and I never find out whether I could have done it better.

Macgilchrist et al. (2020) gave me a frame for this question before I had fully articulated it myself. Writing as social science fiction set in 2040, they offer three competing “histories” of how students and society arrive at that future. In one, “smooth users” improve themselves in pursuit of “frictionless efficiency within a post-democratic frame created by large corporations” (p. 76). In another, “digital nomads,” solopreneurs seeking freedom, individualism, and “aesthetic joy,” exploit state regulations and algorithmic rules while stepping out of the state and deeper into the new economy (p. 76). In a third, more hopeful future, participatory, democratic, and ecological people treat institutions as spaces for exploring more equitable ways of living (p. 76). I am, by trade, a digital nomad in their sense already. The question is whether that way of working survives the very technology it currently depends on.

The Smooth-User Future: When AI Absorbs the Teaching

The most immediately plausible future is the one where my teaching role simply gets absorbed, and the client in my Tuesday vignette becomes the smooth user: the assistant is her frictionless first move, not the exception. Selwyn (2024) argues that the current hype around AI has made it hard to properly weigh its educational costs, and names, among his concerns, “the losses incurred through reorganising education to be more ‘machine readable’” (p. 3). Translate that out of the university and into a small business’s back office: right now, when a client does not understand what their website analytics are telling them, the explanation I give is shaped by what I know about their specific business. If a general-purpose AI assistant can answer the same question competently and instantly, the relationship does not get destroyed so much as it gets reorganized into something more machine-readable, a prompt and a response, instead of a conversation with someone who remembers their business.

The EDUCAUSE (2025) Horizon Report names this risk directly. It warns that free AI tools are becoming so common, and critical data literacy so thin, that just checking whether an AI’s answer is actually true may be one of the defining challenges of the next decade (p. 5). The same report sketches four longer scenarios for 2035, from Collapse to Growth, and AI runs through all of them differently: free and everywhere in the futures it treats as a risk, personalized and abundant in the ones it treats as a win (EDUCAUSE, 2025). Read against my own trade, both framings are describing the same mechanism from opposite ends: free or near-free AI access changes what a person is willing to pay a human for. The report is written about institutions, not solo consultancies, but the mechanism does not stop at the university gate.

The Digital-Nomad Future: Deskilling, or Something Narrower

Here is where Macgilchrist et al.’s (2020) scenario gets uncomfortable for me specifically. Selwyn et al.’s (2020) social-science-fiction exercise, imagining an Australian high school in 2030 through five vignettes, names “digital deskilling” as one effect of increasingly platformized schooling (p. 90). Read onto my own trade, that term almost fits, but not quite. Deskilling, in the sense Braverman (1974) first developed it in Labor and Monopoly Capital, describes management separating the conception of work from its execution, so a worker’s craft knowledge gets appropriated into a system that no longer needs the worker’s judgment to run it. That is not quite what happens to me if an AI assistant answers a client’s question well. My skill does not get appropriated into the tool; it gets duplicated by it, cheaply, over and over. Strictly, this is not deskilling. It is disintermediation: I do not stop being skilled, I stop being the reason anyone needs to ask.

Why I Am Leaning Pessimistic, and Why That Might Be Too Simple

Bozkurt et al.’s (2023) collective reflection, in which thirty-six scholars each wrote two deliberately different speculative vignettes without seeing each other’s entries, in order to surface “the gray areas of optimistic and pessimistic future scenarios” (p. 57), is useful here because it refuses to let either mood stand alone. My own pessimism is narrower than straightforward replacement: AI gets good enough, and cheap enough, that an owner simply asks it first, long before it is actually better at understanding their specific business than I am.

Selwyn (2021) complicates even that narrower pessimism, in a way I had not considered before reading him for this unit. His argument in “Ed-Tech Within Limits” is that the unconstrained growth of digital technology assumes an unconstrained supply of energy and raw materials, an assumption that climate instability will not indefinitely honour. Selwyn does not put this in price terms, but if it holds, it has a price consequence: large, data-intensive AI systems could become more expensive to run, not cheaper, as environmental and resource limits bite. He makes a related point more directly elsewhere, naming the ecological and environmental costs of data-intensive AI as one of his standing concerns (Selwyn, 2024, p. 3). My fear that AI becomes cheap enough to ignore me assumes AI keeps getting cheaper. That assumption, not the technology itself, is the part of my pessimism I am least sure of.

The Collective-Agency Future, and the Honest Limit of My Own Answer

Macgilchrist et al.’s (2020) third future, collective agency, is the one I am least naturally drawn to describe, and reading my own proposed response to disintermediation against it tells me why. Veletsianos et al. (2024), introducing a special collection on higher education futures, note that amid today’s economic, demographic, and ecological uncertainty, many scholars have argued it is time to imagine alternative futures “that are more hopeful, more equitable, and radically more just” (Abstract). My own answer to disintermediation, that the part of my work an AI assistant cannot yet do is being accountable to one specific business, in one specific town, over years, is still an individual answer. It is my version of Macgilchrist et al.’s digital nomad, working the trust-based gap rather than the regulatory one, not a step toward their third future at all. A genuinely collective version would look different: independent consultants sharing what we learn instead of guarding it, or small business owners teaching each other the way I currently teach them one at a time. Polanyi’s (1966) observation, in The Tacit Dimension, that “we can know more than we can tell” (p. 4), is one way of naming what might survive being compressed into a tool either way: the explicit, repeatable part of my explanations is exactly what a language model is good at compressing, but the tacit, situated judgment about one specific business is harder to hand over, whether I keep that judgment to myself or share it with other practitioners. I do not know yet which version of that future I am actually building.

Conclusion

I started this essay asking whether a small business will still need, want, or be able to afford someone like me by 2035. The resource-limits argument (Selwyn, 2021) is what unsettled my own pessimism most: my fear depends on AI getting cheaper indefinitely, and that is the least examined assumption in it. What remains is a narrower, economic worry rather than a technological one. AI does not need to be better than me at my job; it only needs to be cheap and convincing enough that asking it feels like the obvious first move. None of these futures is finished being written, including, in a small and specific way, the one I am currently living in.

AI-Use Plan

I used AI, specifically Claude Code, in methods for this assignment. First, as a research assistant: I had it double-check my work after completing the initial draft, and ensure my research was documented properly (author names, year, journal, page range, DOI) against Crossref, OpenAlex, and, where available, the publisher’s own PDF, rather than trusting my own verification. Second, as a conflicting persona for a final pass of the prose, working from an outline and argument I had already decided (the three-futures structure applied to my own consulting practice, the disintermediation-versus-deskilling distinction, the Selwyn 2021 complication), which I then reviewed and revised.

My guardrails: every citation was checked against actual source metadata or text before being used, not generated from a model’s memory of “a paper like this.” That check caught two real errors in an earlier draft: a reference that put a co-author’s given name in the surname position (Veletsianos et al.’s fourth author is Alhadad, not “Sakinah”), and a wrong page range on the Selwyn (2024) article. Both are fixed in the final reference list below. I did not ask AI to imagine my future for me; the central idea, that my own professional survival is the lens for this assignment, came from my own situation.

References

Bozkurt, A., Xiao, J., Lambert, S., Pazurek, A., Crompton, H., Koseoglu, S., Farrow, R., Bond, M., Nerantzi, C., Honeychurch, S., Bali, M., Dron, J., Mir, K., Stewart, B., Costello, E., Mason, J., Stracke, C. M., Romero-Hall, E., Koutropoulos, A., … Jandrić, P. (2023). Speculative futures on ChatGPT and generative artificial intelligence (AI): A collective reflection from the educational landscape. Asian Journal of Distance Education, 18(1), 53–130. https://doi.org/10.5281/zenodo.7636568

Braverman, H. (1974). Labor and monopoly capital: The degradation of work in the twentieth century. Monthly Review Press.

EDUCAUSE. (2025). 2025 EDUCAUSE Horizon Report: Teaching and learning edition. https://library.educause.edu/resources/2025/5/2025-educause-horizon-report-teaching-and-learning-edition

Macgilchrist, F., Allert, H., & Bruch, A. (2020). Students and society in the 2020s. Three future “histories” of education and technology. Learning, Media and Technology, 45(1), 76–89. https://doi.org/10.1080/17439884.2019.1656235

Polanyi, M. (1966). The tacit dimension. University of Chicago Press.

Selwyn, N. (2021). Ed-Tech within limits: Anticipating educational technology in times of environmental crisis. E-Learning and Digital Media, 18(5), 496–510. https://doi.org/10.1177/20427530211022951

Selwyn, N. (2024). On the limits of artificial intelligence (AI) in education. Nordisk Tidsskrift for Pedagogikk og Kritikk, 10(1), 3–14. https://doi.org/10.23865/ntpk.v10.6062

Selwyn, N., Pangrazio, L., Nemorin, S., & Perrotta, C. (2020). What might the school of 2030 be like? An exercise in social science fiction. Learning, Media and Technology, 45(1), 90–106. https://doi.org/10.1080/17439884.2020.1694944

Veletsianos, G., Houlden, S., Ross, J., Alhadad, S., & Dickson-Deane, C. (2024). Higher education futures at the intersection of justice, hope, and educational technology [Editorial]. International Journal of Educational Technology in Higher Education, 21(1), Article 43. https://doi.org/10.1186/s41239-024-00475-0

Reflection

Using AI this way confirmed something about my own writing process: generating a first draft was the easy part. The hard part was holding seven sources in my head at once and making sure no claim drifted past what the source actually supports, which is exactly where the AI-assisted check earned its place, catching a misattributed author and a wrong page range before they could reach a marker. The part of this process I would not want to automate away is the moment a second reading pushed back on my first idea: Selwyn’s (2021) resource-limits argument complicating my own pessimism, and the realization, only once I checked Selwyn et al.’s (2020) actual term against my own situation, that “disintermediation” fits what I am afraid of better than “deskilling” does. That correction came from checking a source properly, not from asking AI to tell me what to think, which is the distinction this course has been asking me to hold onto since Assignment 1.