If you have any interest in what the future holds for AI, and how the ancient origins of life and our physical world have sculpted our world (from minerals to cognition), then you must read Hayles’ latest work. I suggest that you keep it proudly by your coffee maker and read one precious morsel at a time to cogitate over and to switch your perspectives throughout your day. Here is a short summary in FAQ style to whet your appetite and, of course, the book is so much richer than this! Hayles provides the structure and the details.

What kind of non-human perspectives, umwelts are out there?

First, Hayles is invested in defining cognition beyond our human world perspective—our “umwelts”. Hayles borrows the term “umwelt” from Jakob von Uexküll as a shorthand for looking at the world and springs off this concept to meticulously explore various umwelts: from bacteria to cells, from bees to trees. Bacteria and cells use “micromimesis” to recopy viral genetic materials as a “most wanted” ID database to recognize these intruders (104).¹ Is this cognition? When temperatures drop, deciduous trees go through abscission to withdraw nutrients from leaves and store them in their roots. Is this cognition? A bee, finding a flower, goes back to the hive and wiggles her butt—and other bees go off to that same flower. Is this cognition? Hayles replies with a resounding yes! “[A]ll biological life forms have cognitive capabilities.” (212)

This framework makes sense to me—I cannot know what other humans think, and I infer their cognition based on their reactions to my words—but this uneasy, daily interpretation rests on assumptions and understandings that are probably right but not absolutely guaranteed to be so (155). So, yes, let’s extend the idea of cognition far beyond the human reach (152)—as any cat or dog owner is wont to do (202).

What are these new perspectives and why do we need them?

In her latest work, Hayles further expounds on her interactive cognitive framework (ICF) theory of how a life form creates meaning within their umwelts—their perspectives and life contexts. By expanding our understanding of cognition beyond human umwelts, Hayles thus locates human ICF within an evolutionary cosmos of cognition (29). The more we can turn from our human perspective to explore other perspectives, we can understand our position in a multi-layered, dynamic system of non-human thought. We can shape our thinking to engage with this complex system of interdependencies (symbionts)² to help all of us (human, nonhuman) to survive—even in the face of the consequences of our hum history and actions (231). She then extends the idea of umwelt to AI: “Then we may ask what the computer has knowledge about in its internal milieu—in effect, what the computer’s umwelt is” (67).

Our human umwelt (cf 150) differs radically from that of machines or bacteria or animals. Acknowledging that our human perspective is not the sole perspective in these complex systems leads us to radically new ways of considering complex problems. Setting our human role and perspective within a complex multileveled system helps us shape our thinking to survive and thrive in a universe that is not simply human. We would expand our understanding of social problems beyond human economics—even as complicated levels as the Federal reserve and structural problems such as the fact that there isn’t a planetary oversight body for human impacts like climate change, pollution, etc.— “with a recognition that human lives are enmeshed with, and ultimately dependent on, a wide variety of other lifeforms, from viruses and bacteria to insects, animals and plants, and increasingly in developed societies, on cognitive media as well” (231).

What is the urgency here?

Hayles’ main point is that we need to stop solely seeing cognition through human minds and eyes and expand our view of knowledge far beyond anthropocentrism. Hayles’ urgency for this paradigm shift is twofold: we desperately need these new viewpoints to address the climate crisis and to function within a brave new world that includes such AI (219, my mangling of The Tempest).

For the first urgent path, Hayles urges us to consider ecological relationality, that is, to “shift our vision to seeing humans as guardians who have the responsibility to ensure the welfare of all cognitive beings, biological as well as technical” (219). This new vision would allow us to consider other umwelts to create win-win solutions for all levels of life and cognition.

For the second, we need to forge new visions to understand, analyze, and participate in an AI world (227). We can use a hybrid human-techno system that Hayles terms “cognitive assemblages” to “endorse programs, procedures, and rules that turn the energies of capitalism toward planetary health and sustainability…” (232).

How can we handle AI within a techno-human umwelt?

AI has far surpassed the Turing test at this point (167). The world will no longer turn back to an age before the atomic bomb, and neither will it turn back to an age before AI (unless, of course, the bombs wipe everything out—my comment and not in Hayles). Still, Hayles provides advice and comfort to those struggling to analyze and teach literature (and other fields by extension) after AI has entered the scene. She addresses the citation difficulties with ChatGPT, as the AI works on dynamical processes. Just as we can never step in the same river twice, ChatGPT will never be able to write the exact same answer to the same prompt twice. How can we replicate these texts? What would Gutenberg say? Hayles ponders these questions as well: “Hence citation depends entirely on the assertions of the one who quotes, because they cannot be verified by anyone else. The resulting uncertainties destabilize the whole enterprise of literary criticism, which traditionally has treated exact quotation and citation as the sine qua non for acceptable work” (148). Moreover, AI creativity depends on loosening controls so that AI can refuse commands “based on its own judgment of whether the problem is interesting or not” (176), yet regulations are needed to protect human safety and human rights (177). This reminds me of Asimov’s rules of robotics (which Hayles does not cite).³

Nonetheless, Hayles offers practical advice such as assigning students projects that incorporate AI: show your prompts and how ChatGPT responded, then show what you did with that response; ask ChatGPT for a series of references to articles about a topic and then find those references—if they are not fictional. Hayles analyzes works of fiction like Kim Stanley Robinson’s Ministry for the Future (cf 243) as well as the European Union’s Artificial Intelligence Act (177) for possible political solutions.

How can we apply these perspectives to electronic literature and literature criticism?

But what does this specifically have to do with literature and literary criticism? First, we can use the perennial lit-crit search for meanings that are, as Hayles puts it, “‘really’ in the text as distinct from ones we project onto it” to give us a literary-critical technique to examine the “context-specific” narrative productions of ChatGPT and other AIs (154). It is time to revisit that age-old assumption that the “texts it [lit-crit] interrogates have been written by humans with language processed by the human brain,” as Hayles notes (139). By adding new non-human perspectives, we can completely “shake to the core” traditional methods of inquiry for lit-crit and beyond, to other critical inquiries. What does it mean for an AI to write? Is text output from AI-language models different than human-programmed generative works like Montfort and Strickland’s Sea and Spar Between (258, Note 5)?

Hayles throws down the gauntlet to explore these perspectives of inquiry:

“If, however, literary criticism rises to the challenge, begins to develop strategies that recognize this profound difference, and articulates interpretive techniques appropriate to it, then the productions of LLMs [large language models] can enrich the literary canon, recognized as literary texts worthy of analysis in their own right”(181).

I would love to see this community take up these challenges.

What do we, e-lit folks, need to do now?

Storytelling, particularly the complex stories with multi-layered connections that electronic literature can provide, is critical to Hayles’ urgent missions. We need to write stories so that readers can recognize differences between human and AI umwelts (164). We need to use the power of electronic literature to create complex, deeply intertwingled works to reflect the amazing interweaving of so many different cognitive world perspectives (umwelts). We can then analyze these differences and use them to develop insights for humans living with AI (182)—and again write these stories, for “we will need new fictions to read, different stories to tell, and novel mediating mechanisms to connect us to each other and to our global problems” (221).

Summary

Bacteria to AI delivers on its objective to shift our thinking and embrace many integrative cognitive frameworks to “open the way for more positive futures that nevertheless acknowledge dystopian possibilities” (208). This book is stunningly organized, carefully leveling up from the micro to the global scales. Each invitation to consider the world from a new perspective (from bacteria and xenobots to artificial intelligence) is laid out with a preamble, journeys through that perspective with other scholars and researchers and summarizes with urgent calls for thinking in these new ways.

While the wealth of what is here far overshadows any lacuna I might have wished for, I must confess, however, to a bit of a let down at the final chapter. The book ends, far too abruptly, with an insightful summary of the Ministry for the Future. I was looking forward to an overall summary discussion of how all of these umwelts; how these non-human cognitive perspectives are linked, and how they work as a whole; how we collaborate on perspective levels from the cellular to the universal; and how we, from our human perspective, could work within all of these lenses to survive and thrive. Perhaps this will be Hayles’ next project?

Works Cited

Asimov, Isaac. “Runaround”. ChatGPT summary, 2025. Suggested edition: Asimov, Isaac. “Runaround”, Astounding Fiction, Street and Smith, March 1942, pp. 94-103.⁴

Hayles, N. Katherine Bacteria to AI: Human Futures with Our Nonhuman Symbionts. The University of Chicago Press, 2025.

Footnotes

1. All page numbers are taken from the paperback version, ISBN 13: 978-0-226-83747-5. ↩

2. Symbionts are organism living in a partnership with another organism which Hayles carefully explains with rich examples. And I confess, since reading this book, I see these interdependencies and intertwingled environments everywhere. ↩

3. ChatGPT gave me this summary: The Three Laws of Robotics, devised by Isaac Asimov, are:

   A robot may not injure a human being or, through inaction, allow a human being to come to harm.

   A robot must obey the orders given it by human beings, except where such orders would conflict with the First Law.

   A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

   These laws were first introduced in Asimov’s 1942 short story “Runaround” and have since become a foundational concept in discussions about artificial intelligence and robotics.  ↩

4. Editor’s note: The original Three Rules of Robotics, as originally written by Asimov, are available on p. 100 and read:

   ‘“Now, look, let’s start with the three fundamental Rules of Robotics — the three rules that are built most deeply into a robot’s positronic brain.” In the darkness, his gloved fingers ticked off each point.

   “We have: One, a robot may not injure a human being under any conditions — and, as a corollary, must not permit a human being to be injured because of inaction on his part.”

   “Right!”

   “Two,” continued Powell, “a robot must follow all orders given by qualified human beings as long as they do not conflict with Rule 1.”

   “Right!”

   “Three: a robot must protect his own existence, as long as that does not conflict with Rules I and 2.”

   “Right! Now where are we?” ↩