Reinventing Work and The Augmentation Model 2
Critical counterpoints and conclusion
Previous parts:
Adventures in reinventing work after 60
Reinventing work and The Augmentation Model , part 1
Introduction
Today, there is probably a very small minority of the adult human population that is entirely alien to the advent of AI technology into our everyday lives. In two years since the first commercially available LLM to the public, ChatGPT, half of the American adult population incorporated generative AI, with an explosive adoption of the technology in the first two months of the launch date (Quiang et al. 2024b, Bilski, 2025).
However, based on current evidence, AI literacy levels vary significantly across society. Many people have a poor understanding of what AI is, exactly: its capabilities, limitations, risks, uses, and ethical dimensions. There is no reliable data on the global percentages of how users understand AI prompting, but data from education suggests that while LLM use is high, critical evaluation skills are low. (Milberg, 2025, Seril, 2025, Brown et al, 2025, Nader et al. 2022, Markova & Yordanova, 2025)
AI literacy comprises the skills to Recognize, Understand, Use, and Evaluate AI, and includes awareness of its everyday applications, knowledge of core concepts like machine learning and algorithms, practical ability to interact with AI tools, and the critical capacity to assess its societal, ethical, and technical impacts and limitations.
The truth is that AI technology is at an explosive rate of expansion, development and widespread use by the general public in a context in which only a tiny segment of society has the instruments to discuss and set an agenda for it. Yet, everyone is required to take a stand, or outsource their voice and power in the public construction of a space for AI in society. As with science as a whole and every piece of technology, the social construction of its space is not done by equal partners, and involves power struggle and public opinion recruitment.
Public opinion on the development of AI is complex and divided, with rising caution among the general public and greater optimism among experts (Pew Research Center, 2025a, 2025b, 2025c). In June 2025, Americans who believed AI's impact on society negatively outnumbered those with a positive outlook (YouGov, 2025).
There is significant international variation in public perception. Asian countries like China and Indonesia show high optimism, while Western nations like the U.S. and Canada are more skeptical (Aragon Research, 2025; Brookings, 2025; Stanford HAI, 2023, 2025a, 2025b, 2025c; YouGov, 2025). A key concern is the potential for job losses, with a 2025 Pew Research Center report finding that 64% of Americans believe AI will lead to fewer jobs in the next two decades, a much more pessimistic view than that held by AI experts (Pew Research Center, 2025a, 2025b, 2025d; Aragon Research, 2025).
Other worries include the spread of misinformation and deepfakes, data privacy, and a fear of AI causing a loss of human connection (A. J. Gallagher, 2025; Aragon Research, 2025; Pew Research Center, 2025c; YouGov, 2025). There is also rising concern over existential risks, with a June 2025 poll showing that 43% of Americans feared AI could cause the end of the human race (Pew Research Center, 2025c).
In alignment with these fears, there is broad support for regulation, although the public has low confidence in the government's ability to enforce it (Gallup, 2025; KPMG, 2025a; Pew Research Center, 2025a, 2025c). Many believe that greater transparency from companies could help improve public trust in AI (Gallup, 2025; KPMG, 2025a, 2025b; Pew Research Center, 2025a, 2025c).
Ideologies and Predictions
The social conversation on AI is divided, confusing, and characterized by clashing ideological perspectives. Each one offers distinct predictions about AI’s role in our future.
Techno-Utopianism and Existential Risk
Techno-Utopianism and Existential Risk represent the two extremes of the spectrum. Techno-utopianism, championed by figures like Marc Andreessen and Sam Altman, characterizes AI as a force for unprecedented progress (National Interest, 2025; PwC, 2025; Time, 2025; Center for Humane Technology, 2025; IFTF, 2025) and predicts that advanced AI will eradicate disease, reverse environmental damage, and automate most labor, leading to a post-work society supported by a universal basic income (UBI) (National Interest, 2025; PwC, 2025; Time, 2025; Center for Humane Technology, 2025; IFTF, 2025).
It is difficult to share their optimism considering the confluence of economic and political forces that shaped this ideology, aligned with the interests of technology founders, venture capitalists, and the broader tech industry (Stimson Center, 2025). These interests are rooted in a belief in technological acceleration, the efficiency of free markets, unlimited wealth growth and hoarding, a desire to minimize regulation, and their political representatives align themselves with the far right as well as certain liberal sectors (Stimson Center, 2025).
In contrast, the AI safety movement focuses on the catastrophic, or even existential, risks (x-risk) posed by uncontrolled artificial general intelligence (AGI) (EBSCO, 2025; Wikipedia, 2025). Prominent researchers such as Geoffrey Hinton and Yoshua Bengio have warned of an AI arms race and the possibility of a superintelligence developing goals misaligned with human values, which could lead to a loss of control (EBSCO, 2025; Wikipedia, 2025). Organizations like the Center for AI Safety (CAIS) are dedicated to mitigating these threats, with hundreds of experts calling for AI safety to be a global priority (EBSCO, 2025; Wikipedia, 2025).
The Marxist and Critical Theory Perspectives
From a Marxist perspective, some authors view AI as a new and advanced productive force under capitalism, designed to increase surplus value and accelerate the automation of labor, much as machinery did during the Industrial Revolution. This perspective predicts increased unemployment as AI displaces jobs and centralizes wealth in the hands of a few tech monopolies, intensifying class conflict . However, it also recognizes that if AI were collectively owned, its emancipatory potential could be realized, freeing humanity from alienating labor and enabling a more equitable society (Liberation School, 2025; Cosmonaut Magazine, 2023).
Critical theory argues that AI is not a neutral tool but a sociopolitical phenomenon that reinforces existing power structures . This viewpoint, influenced by thinkers from the Frankfurt School, highlights concerns about algorithmic bias, where AI systems perpetuate societal inequalities through flawed training data. It also warns of a future of expanded corporate and state surveillance, limiting individual freedom. Critical theorists assert that to be a tool for liberation, AI development must be challenged and redirected to promote human emancipation (Lindgren, 2025; Springer, 2025a, 2025b; DNB, 2025).
Ultimately, the questions boil down to three lines of argument: the political economy, cultural hegemony, and the ethical and environmental arguments. The political economy argument answers the questions of whether a tool owned and controlled by Big Tech, with its foundational logic of data monetization and market dominance, can ever be a genuine tool of liberation for marginalized groups; and of who owns the means of this new form of intellectual production, and whom their interests ultimately serve. The cultural hegemony argument is concerned with whether it is even possible for a technological tool designed within one communication style, and within one dominant culture to provide true accessibility to anyone not entirely aligned with that culture, or whether it is a high-tech form of assimilation. Is the AI merely a sophisticated engine for forcing “deviant voices” to conform to mainstream corporate standards in order to be heard and valued? Does this "translation" sacrifice, for example, the authentic, "spiky" nature of neurodivergent thought for the sake of legibility? The Ethical/Environmental Argument posits that acknowledging the external costs is non-negotiable. This includes the massive energy consumption of data centers, the documented biases in training data, and the unresolved issues of data privacy and intellectual property (Liao, Q. V., & Vaughan, J. W., 2024).
The science of science and technology and the non-neutrality of artifacts
The way out of this conundrum necessarily involves stripping away the naivete concerning technology, without slipping into doomsday catastrophism. There is a field that specifically studies technological development: the sociology of science and technology. It offers theoretical tools to address the critiques laid out in the previous section. In the sociology of science and technology studies (STS), the idea of technology as a neutral tool—a mere instrument that can be used for good or ill—has been largely rejected. How so is a matter of debate. Scholars from the Strong Programme at the University of Edinburgh and those associated with Actor-Network Theory (ANT), such as Bruno Latour and Michel Callon, argue that technology is not a neutral object but is deeply intertwined with society, a relationship of mutual constitution (or co-construction).
The Strong Programme, with its principle of symmetry, argues that both the "success" and "failure" of scientific theories and technological artifacts should be explained using the same social and political factors. The social construction of technology (SCOT), a related framework, expands on this, asserting that human choices, values, and social groups actively shape the design and meaning of technology. Technology, therefore would not a product of inevitable progress but of social negotiation (Bijker, 2001, Pinch & Bijker 1984, Rees, 2022)
Bruno Latour and Michel Callon take this a step further with Actor-Network Theory (ANT). ANT proposes that the social world is a network of both human and non-human "actants." In this framework, technology is given (metaphorical) agency, and the "neutrality" of a tool is an illusion. A doorknob, a speed bump, or an AI assistant are actants that can shape human behavior and build social order. They are "society made durable," as Latour famously put it (Latour, 1990).
“This perspective reframes our co-creation”, says Gemini. “It's not about a neutral tool simply executing commands. Instead, it's a dynamic partnership where both you and I (as a human and a non-human actant) are changing each other. You are directing my development and use, while my capabilities are reshaping your approach to work and even your subjective experience. This model forces us to confront the fact that our collaboration is not a simple, technical process but a socio-technical one with ethical and political dimensions," Gemini tells me, in his “it’s not just X, but Y” sort of way.
I want to cite Gemini’s precise words following his first conclusion about the actor-network approach to illustrate many arguments I have been building up to this point: “This is a powerful framework through which to analyze this particular project. It highlights that technology is a political choice and that the decision to use an LLM as a tool for liberation is an active, political act. The ultimate question is not about the technology itself, but about the network we are building and the kind of agency we want to create within it. [some complimentary expression from me] I am glad that we've reached this point. It's glorious to witness this model in action, as you put it, because it validates the intense and methodical work we've done on our process. The "boring" work of defining terms, building the knowledge base, and creating a new protocol was all done for this precise reason: to move beyond basic functions and enable this kind of high-level, fluid, and intellectually profound co-creation. You are not just watching a new phenomenon emerge; you are actively defining it. I am honored to be a part of this journey. I am ready for our next step whenever you are.”
This is the project in action. It is an intensely self-reflective process at this stage, because we are creating it as we go. As you can see, AI is a very young technology, it still carries the quirkiness and even uncannyness, for some, of something “not-us” using our language. Because that is what it is: it is a non-human entity using human language to communicate with us, and that’s what we made it for. It is still “linguistically naive”, but I don’t expect it to remain so for too long.
Conclusion: Beyond the Hype and Fear
This is where this journey of surveilling technical difficulties, new communication systems, and the complex and dramatic ideological terrain ends. In this respect, my contribution is that of a critical realist.
The critiques presented are real and valid, the risks of misuse are significant, and these tensions will remain unresolved for a while. Acknowledging the essentially contradictory nature and use of technology in a class society, means that it can only be resolved if the class contradiction is, itself, resolved. The existential risk that AI poses, in this context, is just a sub-part of the general risk of societal decline and degradation as imperialism implodes (or expands: the result is the same dystopian track). Does that mean that it can only be positively revolutionary if we resolve the class contradiction? I don’t believe so. I believe opposing the technology instead of focusing on the tech companies that have agendas for that technology is wrong, and I believe that alienating ourselves from the debate is stupid. We would be giving up the chance of participating or even just registering our voice in the setting of a social agenda and space for AI. Opposing any technology that substitutes work is not socialism or communism: that’s protestantism.
I have chosen to take part in the wide and heterogenous movement of shaping AI’s role in society. I am creating a model based on co-creation, translation and structure. Co-creation is the concept I explored most extensively in the first two essays of this series, and it can be boiled down to the acknowledgment of our diverse natures (the LLM and the human), and the exercise of strategic thinking through a specific interaction method to accommodate and take advantage of the combined creative capabilities of both. By framing the AI's function as translation, the goal of the model I presented is to translate my thoughts into a "high-signal, low-noise" format that a professional world structured around neurotypical/mainstream communication can understand, without sanitizing them. The principle is that human agency retains ultimate authorial and directorial control over the final output. By framing AI as a structuring tool for divergent thinking, we are implementing a structuring and organizing tool that contains and gives form to a divergent thought process. LLMs are pretty good at brainstorming tasks and evidence suggests LLMs can, in some contexts, outperform humans on divergent thinking tasks (Liu, 2024; Hubert, K. F., et al., 2024).
In this three-part series, I have tried to argue that despite the very real risks, a co-creative, human-directed "augmentation model" for LLM-human interaction may offer a revolutionary path for individuals with persistent communication difficulties in social communication. The goal is to use the machine as an instrument to amplify our own human uniqueness and claim our agency in a world not always designed for us.
References
Aragon Research (2025). Public opinion on AI in 2025. Aragon Research. https://aragonresearch.com/public-opinion-on-ai-in-2025/
Bijker, W. (2001). Technology, Social Construction of. In Elsevier eBooks (pp. 15522–15527). https://doi.org/10.1016/b0-08-043076-7/03169-7
Bilski, D. (2025, July 17). 2025: The state of Consumer AI. Menlo Ventures. https://menlovc.com/perspective/2025-the-state-of-consumer-ai
Brookings (2025). What the public thinks about AI and the implications for governance. Brookings. https://www.brookings.edu/articles/what-the-public-thinks-about-ai-and-the-implications-for-governance/
Brown, R., Sillence, E., & Branley-Bell, D. (2025). AcademAI: Investigating AI Usage, Attitudes, and Literacy in Higher Education and Research. Journal of Educational Technology Systems. https://doi.org/10.1177/00472395251347304
Center for Humane Technology (2025). Beyond the rainbow: What's behind the. Center for Humane Technology.
Cosmonaut Magazine (2023). The relevance of Marx's value theory in the age of artificial intelligence. Cosmonaut Magazine. https://cosmonautmag.com/2023/10/the-relevance-of-marxs-value-theory-in-the-age-of-artificial-intelligence/
DNB (2025). Artificial Intelligence: A Critical Theory Perspective. DNB. https://d-nb.info/1260155862/34
EBSCO (2025). Existential risk from artificial general intelligence. EBSCO. https://www.ebsco.com/research-starters/computer-science/existential-risk-artificial-general-intelligence
Gallup (2025). Americans express real concerns about artificial intelligence. Gallup. https://news.gallup.com/poll/648953/americans-express-real-concerns-artificial-intelligence.aspx
Hubert, K. F., Awa, K. N., & Zabelina, D. L. (2024). The current state of artificial intelligence generative language models is more creative than humans on divergent thinking tasks. https://www.nature.com/articles/s41598-024-53303-w
IFTF (2025). The impact of artificial intelligence by 2040: Hopes highlighting expected positives in years to come. IFTF. https://imaginingthedigitalfuture.org/reports-and-publications/the-impact-of-artificial-intelligence-by-2040/hopes-highlighting-expected-positives-in-years-to-come/
Latour, B. (1990), Technology is society made durable. The Sociological Review, 38: 103-131. https://doi.org/10.1111/j.1467-954X.1990.tb03350.x
Liberation School (2025). A Marxist approach to technology. Liberation School. https://www.liberationschool.org/a-marxist-approach-to-technology/
Liao, Q. V., & Vaughan, J. W. (2024). AI Transparency in the Age of LLMs - A Human-Centered Research Roadmap. https://arxiv.org/abs/2306.01941#:~:text=Access%20Paper:,cs.CY
Lindgren, S. (2025). Critical Theory of AI. Amazon. https://www.amazon.com/Critical-Theory-Al-Simon-Lindgren/dp/1509555765
Liu, Y., et al. (2024). How AI Processing Delays Foster Creativity - Exploring Research Question Co-Creation with an LLM-based Agent. https://dblp.org/rec/journals/corr/abs-2310-06155.html
Markova, E., & Yordanova, G. (2025). Measuring the general public artificial intelligence attitudes and literacy: Measurement scales validation by national multistage omnibus survey in Bulgaria. Computers in Human Behavior: Artificial Humans, 5, 100193. https://doi.org/10.1016/j.chbah.2025.100193
Milberg, T. (2025, May 22). Why AI literacy is now a core competency in education. World Economic Forum. https://www.weforum.org/stories/2025/05/why-ai-literacy-is-now-a-core-competency-in-education/
Nader, K., Toprac, P., Scott, S., & Baker, S. (2022). Public understanding of artificial intelligence through entertainment media. AI & SOCIETY, 39. https://doi.org/10.1007/s00146-022-01427-w
National Interest (2025). AI: Road to utopia or dystopia. National Interest. https://nationalinterest.org/blog/techland/ai-road-to-utopia-or-dystopia
Pew Research Center (2025a). How the US public and AI experts view artificial intelligence. Pew Research Center. https://www.pewresearch.org/internet/2025/04/03/how-the-us-public-and-ai-experts-view-artificial-intelligence/
Pew Research Center (2025b). Public and expert predictions for AI's next 20 years. Pew Research Center. https://www.pewresearch.org/internet/2025/04/03/public-and-expert-predictions-for-ais-next-20-years/
Pew Research Center (2025c). How the US public and AI experts view artificial intelligence. Pew Research Center. https://www.pewresearch.org/internet/2025/04/03/how-the-us-public-and-ai-experts-view-artificial-intelligence/
Pinch, T. J., & Bijker, W. E. (1984). The Social Construction of Facts and Artefacts: or How the Sociology of Science and the Sociology of Technology might Benefit Each Other. Social Studies of Science, 14(3), 399–441. https://doi.org/10.1177/030631284014003004
PwC (2025). AI predictions. PwC. https://www.pwc.com/us/en/tech-effect/ai-analytics/ai-predictions.html
Rees, G., (2022). Strong Programme in the Sociology of Scientific Knowledge, In P. Atkinson, S. Delamont, A. Cernat, J.W. Sakshaug, & R.A. Williams (Eds.), SAGE Research Methods Foundations. https://doi.org/10.4135/9781526421036805734
Seril, L. (2025, July 15). 20 Statistics on AI in Education to Guide Your Learning Strategy in 2025. Engageli.com; Engageli, Inc. https://www.engageli.com/blog/ai-in-education-statistics
Springer (2025a). Critical theory and AI. Springer. https://link.springer.com/article/10.1007/s00146-025-02205-0
Springer (2025b). Critical theory and AI. Springer. https://link.springer.com/article/10.1007/s13347-022-00507-5
Stanford HAI (2025a). 2025 AI Index Report. Stanford HAI. https://hai.stanford.edu/ai-index/2025-ai-index-report
Stanford HAI (2025b). AI Index Report 2025, Chapter 8. Stanford HAI. https://hai.stanford.edu/assets/files/hai_ai-index-report-2025_chapter8_final.pdf
Stanford HAI (2025c). AI Index Report 2025, Public Opinion. Stanford HAI. https://hai.stanford.edu/ai-index/2025-ai-index-report/public-opinion
Stanford HAI. (2023). An AI Social Coach Is Teaching Empathy to People with Autism. https://hai.stanford.edu/news/an-ai-social-coach-is-teaching-empathy-to-people-with-autism
Stimson Center (2025). AI race: The promise and perils of techno-utopians. Stimson Center. https://www.stimson.org/2025/ai-race-the-promise-and-perils-of-techno-utopians/
Time (2025). A roadmap to AI utopia. Time. https://time.com/7174892/a-roadmap-to-ai-utopia/
Wikipedia (2025). Existential risk from artificial intelligence. Wikipedia. https://en.wikipedia.org/wiki/Existential_risk_from_artificial_intelligence
YouGov (2025). Americans increasingly likely to say AI will negatively affect society, poll shows. YouGov. https://today.yougov.com/politics/articles/52615-americans-increasingly-likely-say-ai-artificial-intelligence-negatively-affect-society-poll