Communicative AI: A Critical Introduction to Large Language Models

by Mark Coeckelbergh, David J. Gunkel

Large Language Models (LLMs), like OpenAI’s ChatGPT and Google’s LaMDA, are not only the most disruptive and controversial technologies of our time, but also offer an unprecedented opportunity to examine human cognition and philosophically question the very nature of language, communication, and intelligence. What is consciousness? What is language? Are LLMs authors? Are LLMs the end of writing as we know it?

In Communicative AI, Mark Coeckelbergh and David J. Gunkel offer a critical introduction to LLMs, investigating the philosophical significance of this technology and its practical ramifications. Mobilizing resources from contemporary philosophy, history of ideas, linguistics, and communication theory, the book invites us to re-think some long-standing philosophical issues concerning language, consciousness, truth, authorship, and writing. Through a blend of theoretical analysis, accessible explanations, and practical examples, the book provides readers with a comprehensive overview of the role that this powerful new technology is already playing in our lives.

This is a must-read for students and scholars across the humanities and the social sciences, as well as for anyone intrigued by the intersection of technology, language, and human thought.

Genres: Artificial Intelligence, Nonfiction

144 pages, Hardcover

Published July 15, 2025

About the author

Mark Coeckelbergh is Professor of Philosophy of Media and Technology at the University of Vienna. He is the author of New Romantic Cyborgs: Romanticism, Information Technology, and the End of the Machine, AI Ethics (both published by the MIT Press), Introduction to Philosophy of Technology, and other books.

Reviews

[Luc Robinet]

I certainly don’t know how to rate this, much of the proposed thoughts and conclusions *do* make sense. If we take deconstruction and Derrida to their utmost implications and apply that to LLMs, then yes, LLMs and 'AI‘ accelerate this gesture of de-masking written language, and logocentrism in general.

What I have a hard time grasping and accepting is, doesn’t this whole Spiel rather show the ultimate perverse conclusion of deconstruction? In that case the book is less about LLMs in general, but more about the ultimate implications of derrida‘s 'disciples' shown through their handling of LLMs. In that case it morphs into some kind of accelerationist half-stepping, where the intent to accelerate these processes is hidden, maybe even unconsciously proposed - I don’t think that Coeckelbergh and Gunkel consider themselves as accelerationists, but for me it ultimately falls into that current.

Nonetheless, it opens interesting questions; if we follow this rift, which was opened by the authors, we must rather think about different modes of understanding certain texts, and their intention. Not all texts are created equal, and in case of LLMs, they’ve probably not been created at all. They’ve been calculated. What does this imply then?

# - # - # - # - #

Update after second read through: There also seems to be some sort of disdain, or at least lack of understanding when it comes to the aesthetic properties of language and LLMs. Gunkels favorite comparison of "LLMs being akin to sampling, DJs and Remixers" is in this case a relatively limping one. On one hand it shows disregard to the technicalities and the processes of sampling (the end product is not being generated by probabilities; a producer samples something by evaluating the aesthetic content, and further works with those contents to create another work — which in the whole process is driven by their own aesthetic judgement). On the other hand even a remix, DJ Set or what not, is exemplified by the creators own style and manner in the enactment of their work. One of the reasons why people tend to be sceptical of LLMs (or AI in general) is because of it's lack of aesthetic value and creativity. It does open the discussion between creativity and probability though; can mathematical probability be considered a creative endeavor? If yes, then we need to redefine what the properties of aesthetic and creative undertakings entail (which on the other hand, would completely eradicate the human necessity of aesthetic judgement when thinking of art).

Applied to language I see how this supposed deconstruction decentralizes the human element in languages and writing, BUT to kill the author also means to kill the origin of it's style in which said writing appears; to kill the manner in which said writing, or it's language, is written. For me at least, the constitutive difference in approaching this topic would be the question: Are we using technology as a crutch, or are we using technology as a prothesis? A prothesis implies that we have lost something integral to our functioning, and here I sense the danger of expecting a coming deprivation, just to justify the substitution après coup.

[Mikael Hall]

Detta var en bra introduktion och jag gillar greppet som kommer mot slutet av boken att använda post-strukturalistiska tänkare för att tänka kring LLMer. Det sagt är boken alldeles för optimistisk i sin syn på LLMer. Jag tror LLMer kommer vara del i en omvälvande och destruktiv process där vi kommer se en spaltning mellan en elitär människokultur och pöbelns AI-kultur. En process som redan är på gång på många sätt.

[Amy]

An interesting look at the philosophical issues raised by LLMs. The discussion on the deconstruction of the speech/writing binary is interesting, but misses what other scholars in literary studies, literacy education, and social constructivism might have to say. I’m left thinking several key things do not get addressed in this book: languages being social, genre and modality impacting the message, and LLMs’ concerning sycophancy.

[Dewi]

boo 👎
too much philosophy, boring
worse than project hail mary

[Joel]

A helpful book for thinking about AI in the larger context of what writing and language are and what they mean with or without a human 'author.'

The authors are, I believe, philosophers, so I couldn't quite connect with some of their points about, e.g., Plato. I didn't really agree with their conclusions (basically that 'writing ' can exist without a human writer and that's OK for various reasons), though I did find their premises intriguing and reasonable.

It ticks me off that it took these stupid tech companies to finally force me to get into philosophy of language, but I guess I'm going to have to.

[Wade Winter]

I read this for my AI in writing and communication course as the main textbook. It covered a wide breadth of how LLMs and writing interminable, but lacked depth sometimes. This book is good as a textbook because it provides the baseline and framework for me to take some ideas and form an inquiry into AI in writing. I disagreed with the author's conclusion, but their analysis was still thought provoking.