Anthony J. Pennings, PhD

WRITINGS ON AI POLICY, DIGITAL ECONOMICS, ENERGY STRATEGIES, AND GLOBAL E-COMMERCE

The Historical Transitions from Human Expertise to AI Techno-Epistemologies

Posted on | July 8, 2026 | No Comments

Citation APA (7th Edition)

Pennings, A.J. (2026, Jul 08) The Historical Transitions from Human Expertise to AI Techno-Epistemologies. apennings.com https://apennings.com/social_media/the-historical-transitions-from-human-expertise-to-ai-epistemologies/

Introduction

Viewed historically, each information and communication technology (ICT) extends rather than replaces the previous one. Viewed historically, ICT have not displaced one another so much as remediated and extended the symbolic and computational capacities of earlier media. Each new medium preserves the major functions of its predecessors while introducing a new dominant logic that expands the scale, speed, and sophistication of organizational coordination.[1]

Historical Transitions

The sequence begins with writing, whose primary contribution was the ability to record knowledge across time and space. Writing transformed speech into durable symbols, allowing organizations, governments, and civilizations to preserve memory beyond the limits of individual cognition. Laws, contracts, inventories, and correspondence became stable records that could coordinate human activity across generations.[2]

Building upon writing, accounting introduced a new logic of classification. Rather than simply recording events, accounting organized them into standardized categories such as assets, liabilities, revenues, expenses, and inventories. This classificatory grammar made economic activity comparable and governable. Merchants, states, and later corporations could systematically monitor resources, evaluate performance, and coordinate increasingly complex organizations through common symbolic categories.[3]

The digital spreadsheet transformed accounting once again by introducing explicit symbolic computation. Rows, columns, cells, formulas, and functions converted static accounting records into dynamic computational models. Relationships that had once existed only in the analyst’s reasoning became encoded directly into formulas that could instantly recalculate entire financial systems. The spreadsheet therefore shifted organizations from simply recording and classifying reality to actively modeling alternative futures through “what-if” analysis, forecasting, budgeting, and financial simulation.[4]

The emergence of the database extended spreadsheet logic into persistent organizational memory. Databases emphasized storage and retrieval, allowing enormous quantities of structured information to be maintained, queried, and shared across distributed organizations. While spreadsheets remained powerful analytical tools, databases became the authoritative repositories supporting enterprise systems, customer records, logistics, and financial transactions. They separated long-term information management from analytical computation while preserving the spreadsheet’s underlying symbolic grammar.

Machine learning introduced another important transition by replacing explicit formulas with statistical inference. Instead of requiring analysts to specify mathematical relationships manually, machine-learning algorithms estimate those relationships directly from data. Human-authored formulas give way to learned parameters that continually adjust through optimization. The focus shifts from representing known relationships to discovering previously unknown patterns, enabling prediction, classification, anomaly detection, and adaptive decision support.

Large language models (LLMs) extend machine learning beyond numerical prediction into semantic generation and reasoning. Trained on vast corpora of text, code, images, and other symbolic media, they recognize relationships among concepts, synthesize knowledge, generate coherent language, and assist human reasoning across diverse domains.

They remediate not only spreadsheets and databases but also books, libraries, encyclopedias, search engines, programming languages, and organizational documents into a unified semantic environment capable of interacting through natural language.

The next transition is what I describe as SACT-AI. Rather than focusing solely on representation, storage, prediction, or language generation, SACT-AI introduces the dominant logic of coordination. Drawing upon the symbolic grammar established by writing, accounting, spreadsheets, databases, and machine learning, SACT-AI integrates these capabilities into a globally distributed infrastructure that continuously monitors, evaluates, and coordinates economic activity. Within this architecture, symbolic representations become operational. Artificial intelligence no longer simply analyzes information or predicts outcomes; it coordinates liquidity, telecommunications, logistics, governance, and organizational processes across synchronized digital networks.

This historical progression illustrates a broader transition in the role of symbolic media. Writing enabled societies to remember. Accounting enabled them to organize. Spreadsheets enabled them to compute. Databases enabled them to store and retrieve. Machine learning enabled them to infer. Large language models enabled them to reason semantically. SACT-AI extends this trajectory by enabling institutions to coordinate distributed economic activity continuously across time and space.

This is the central historical argument. The future of artificial intelligence is not a break with spreadsheet logic but its operational extension. Spreadsheet logic remains the symbolic and accounting grammar inherited from centuries of organizational practice. AI remediates that grammar through statistical learning and semantic reasoning, while SACT-AI transforms it into an infrastructure for operative mediation in which representation, inference, and coordination become integrated within a globally synchronized system. The defining question for the next generation of information systems is therefore no longer simply how organizations compute knowledge, but how they coordinate action through shared symbolic and computational infrastructures.

References

Bolter, J. D., & Grusin, R. (1999). Remediation: Understanding New Media. MIT Press.
Giddens, A. (1984). The Constitution of Society: Outline of the Theory of Structuration. University of California Press.
Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press.
Grossman, Tomas A.; Mehrotra, Vijay; and Özlük, Özgür (2007) “Lessons from Mission-Critical Spreadsheets,” Communications of the Association for Information Systems: Vol. 20 , Article 60. DOI: 10.17705/1CAIS.02060
https://aisel.aisnet.org/cais/vol20/iss1/60
McCarthy, J. (2007). What is artificial intelligence? Stanford University. (Original work written 2004).
Goody, J. (1986) The Logic of Writing and the Organization of Society. Studies in Literacy, Family, Culture and the State. (Cambridge: Cambridge University Press).
Levy, S. (1989) “A Spreadsheet Way of Knowledge,” in Computers in the Human Context: Information Technology, Productivity, and People. Tom Forester (ed.) (Oxford: Basil Blackwell). He gave me permission to put the article on my website when I was at NYU.
McLuhan, M. (1964). Understanding Media: The Extensions of Man. McGraw-Hill.
Mitchell, T. M. (1997). Machine Learning. McGraw-Hill.
Orlikowski, W. J. (2000). Using technology and constituting structures: A practice lens for studying technology in organizations. Organization Science, 11(4), 404–428.
Poovey, M. (1998). A History of Modern Fact: Problems of Knowledge in the Sciences of Wealth and Society. University of Chicago Press.
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). Learning representations by back-propagating errors. Nature, 323(6088), 533–536.

Notes

[1] Remediation is a constant theme in my work on Spreadsheet Logic and AI because it follows a tradition of media interrogation that focuses on how newer media incorporate older media. What are its implications for epistemology? For the production of power? Newer technologies influence on society such as television, which integrated radio and film.
[2] For an in-depth analysis of writing I liked Giddens’ use of Goody, J. (1986) The Logic of Writing and the Organization of Society. Studies in Literacy, Family, Culture and the State. (Cambridge: Cambridge University Press).
[3] Mary Poovey’s A History of Modern Fact offers one of the most penetrating accounts of how modern economic knowledge came to appear objective, factual, and politically neutral. Poovey, M. (1993) “Figures of Arithematic, Figures of Speech: The Discourse of Statistics in the 1830’s,” Critical Inquiry. Winter, Vol. 19, No. 2.
[4] In an extraordinary chapter Steven Levy’s Levy, S. (1989) “A Spreadsheet Way of Knowledge,” in Computers in the
Human Context: Information Technology, Productivity, and People. Tom Forester (ed.) (Oxford: Basil Blackwell).
AI Prompt(s) What is difference between spreadsheet models and the machine-learning models?

© ALL RIGHTS RESERVED

Not to be considered financial advice. AI is often used, and results are thoroughly interrogated. Links are used for some citations.



AnthonybwAnthony J. Pennings, PhD is a Professor at the Department of Technology and Society, State University of New York, Korea and a Research Professor for Stony Brook University. He teaches AI and broadband policy. From 2002-2012 he taught digital economics and information systems management at New York University. He also taught in the Digital Media MBA at St. Edwards University in Austin, Texas, where he lives when not in Korea.

Comments

Comments are closed.

  • Categories

  • Referencing this Material

    Copyrights apply to all materials on this blog but fair use conditions allow limited use of ideas and quotations. Please cite the permalinks of the articles/posts.
    Citing a post in APA style would look like:
    Pennings, A. (2015, April 17). Diffusion and the Five Characteristics of Innovation Adoption. Retrieved from https://apennings.com/characteristics-of-digital-media/diffusion-and-the-five-characteristics-of-innovation-adoption/
    MLA style citation would look like: "Diffusion and the Five Characteristics of Innovation Adoption." Anthony J. Pennings, PhD. Web. 18 June 2015. The date would be the day you accessed the information. View the Writing Criteria link at the top of this page to link to an online APA reference manual.

  • About Me

    Professor (full) at State University of New York (SUNY) Korea since 2016. Research Professor for Stony Brook University. Moved to Austin, Texas in August 2012 to join the Digital Media Management program at St. Edwards University. Spent the previous decade on the faculty at New York University teaching and researching information systems, digital economics, and global political economy

    You can reach me at:

    anthony.pennings@gmail.com
    apennings70@gmail.com
    anthony.pennings@sunykorea.ac.kr

    Follow apennings on X

  • About me

  • Writings by Category

  • Flag Counter
  • Pages

  • Calendar

    November 2026
    M T W T F S S
     1
    2345678
    9101112131415
    16171819202122
    23242526272829
    30  
  • Disclaimer

    The opinions expressed here do not necessarily reflect the views of my employers, past or present. Articles are not meant as financial advice.

  • Verified by MonsterInsights