Bacteria to AI: Human Futures with Our Non-Human Symbionts

The Italian edition of N. Katherine Hayles’s latest book, Bacteria to AI: Human Futures with Our Non-Human Symbionts (University of Chicago Press, 2025), has just been released by PensaMultimedia. The Italian translation was edited by PRESERVE partners Antonio Carnevale and Gabriele Nino (University of Bari Aldo Moro), and was published with the contribution of the PRESERVE project.

The volume proposes an “Integrated Cognitive Framework” for understanding and comparing cognition across humans, non-human animals, and computational systems. This approach is explicitly designed to unsettle inherited anthropocentrism and to re-situate “the human” within a broader field of meaning-making agents.

From Cognition Beyond Consciousness to Technosymbiosis

The starting point is a simple but radical move: separating cognition from consciousness. For millennia, the two were treated as near-synonyms. Hayles reverses the dependency: non-conscious cognition – the vast array of interpretive processes beneath awareness – is not merely a support for consciousness, indeed it makes it possible. Once this is granted, the scope of cognition expands beyond the human to the entire biosphere. In the natural world, organisms such as bacteria, plants, fungi, and insects interpret information in ways that are essential for their survival. Their behaviors are not merely mechanical reflexes, they are meaningful, biologically grounded responses. As a result, humanity becomes one species among many that participate in distributed meaning-making across overlapping yet distinct Umwelten.

The relocation of cognition in this manner let the author develop the Integrated Cognitive Framework (ICF), a system that maintains visibility of differences while facilitating comparison. “Integrated” here means that the model compares and juxtaposes three different origin stories – from quantum mechanics, evolutionary biology, and technics – to demonstrate a shared emphasis on relationality. Each of these concepts is explained through the lens of interdependence: quantum entanglement undermines isolation; evolution binds organisms to ecologies; and technical infrastructures only gain capacity in networks of sensors, actuators, and computation. When positioned in juxtaposition, these accounts delineate cognition as the interpretation of information within contexts that connect it to meaning.

In which areas are cognition taking place today?

This is becoming increasingly evident: in various assemblages. Modern infrastructures such as transport, logistics, finance and, healthcare, rely on ensembles where humans are indispensable yet not sole decision-makers. The system first processes information through its various sensors, filters, and pre-interpreters. Then, it is approved or overridden by the humans who operate it. Finally, the electromechanical systems enact the desired actions. Human agency does not disappear; it is simply distributed differently. What appears to be a solitary choice is frequently the result of intricate computational, biological, and environmental processes. The assertion that humans make all decisions is empirically inaccurate, and the claim that they make no decisions at all is equally erroneous. The ICF proposes a middle ground, acknowledging the heterogeneity of neuronal, cellular, and silicon embodiment without presuming their comparability.

From this perspective, the author coins the concept of technosymbiosis in order to depict this relationship. Societies have long utilized non-human forces, such as fire, steam, and fossil energy. However, the current age of computing has us working closely with non-human entities that are not only powerful but also cognitive. Our cars, factories, and networks are designed to not only execute but also sense, predict, and coordinate. This intimacy brings both capability and vulnerability. A systemic digital failure would have far-reaching consequences, potentially paralyzing critical sectors such as mobility, energy, finance, and communication. This underscores the crucial nature of these systems, highlighting that they are not mere optional accessories but rather the fundamental conditions for social life to function effectively. The term “symbiosis” is a biological term that encompasses a variety of relationships, including parasitism, commensalism, and mutualism. It does not carry a romantic connotation rather; it asserts the co-constitution that our identities are shaped by the assemblages we form. This reframing carries direct ethical and governance implications. If cognition and agency are distributed, responsibility cannot reside in a single node. For law enforcement and public bodies, this has implications for evidence chains, model evaluation, and oversight. The definition of “explanation” is affected when interpretation is pre-filtered by a device. The concept of “control” is also impacted, as safety relies on real-time interaction between operators and predictive systems.

The ICF also guards against two symmetrical errors in AI discourse. The first is hybris – the belief that humans remain fully “in control” even as their judgments are influenced by algorithmic defaults. The second is abdication – that “the system decided,” as though computation possessed alien will. The framework calls for a detailed description of the systems’ functionality, including their interpretation capabilities, optimization processes, and the management of uncertainty. These descriptions must be aligned with legal and ethical commitments. Cognitive assemblages should not be used as justifications as they are the terrain where responsibility is redrawn.

Why this is relevant in PRESERVE context

Hayles’s Integrated Cognitive Framework is highly resonant with PRESERVE’s ambition to build an ethical, privacy-preserving decision-support system for European law enforcement authorities. PRESERVE does not simply add another AI tool to existing practices: it creates a complex cognitive assemblage in which social media streams, deep and dark web data, police databases, federated learning infrastructures and human analysts jointly interpret signals of cybercrime and terrorism. Hayles’ notion of technosymbiosis offers a vocabulary for thinking about this hybrid ecology, where human and non-human agents co-produce meaning, risk assessments and investigative priorities.

This perspective is particularly valuable for PRESERVE’s focus on bias mitigation, explainability and human-rights compliance. By treating cognition and agency as distributed, the framework helps to avoid both the fiction of full human control and the opposite temptation to blame “the system” alone. It supports the project’s effort to design accountability chains, evidence standards and training activities that acknowledge how decisions emerge from interactions between LEAs, data pipelines and AI models. In this sense, Bacteria to AI is not only a theoretical reference, but a conceptual toolkit for designing, governing and critically reflecting on the privacy-preserving AI platform that PRESERVE aims to deliver.

Conclusions

By centering non-human cognition, the book moves beyond the binary of human versus machine. It emphasizes the continuity between cellular and computational interpretation without oversimplifying the distinction between the two, providing a set of tools – the ICF, cognitive assemblages, technosymbiosis, for thinking across biological and technical scales. The practical lesson is clear: rather than denying our dependencies, we must cultivate them wisely. We must design for resilience, legibility, and justice in the hybrid systems we already inhabit. If the future is shaped by non-human symbionts, the ethical question is not whether to partner, but how. The ICF asks us to take that “how” seriously.