When metaphors become models
The article begins with an observation that seems obvious yet carries profound consequences: artificial intelligence and brain science now use overlapping vocabularies. Terms like “learning,” “memory,” “attention,” or “perception” have migrated from cognitive psychology into computer science, while concepts such as “computation,” “information processing,” and “architecture” have made the reverse journey, shaping how neuroscientists talk about the human brain.
Floridi and Nobre (2024) show that this mutual borrowing is not neutral. Metaphors, once introduced, often solidify into models that guide research, funding, and even public imagination. When we describe an algorithm as “intelligent,” we do more than simplify language: we endorse a worldview where cognition is computation, and where human and artificial systems are placed on a continuum.
The double-edged power of cross-disciplinary language
This crosswiring, the authors argue, has both benefits and risks. On one side, it has enabled genuine innovation: the exchange between disciplines has generated powerful tools, from artificial neural networks inspired by biology to computational models that help neuroscientists simulate brain activity. Without this shared language, much of modern AI would not exist.
On the other side, linguistic blending can distort understanding. When scientists or the public conflate “learning” in machines with learning in humans, they risk overlooking what makes cognition distinctively human: context, meaning, emotion, and intentionality. A “neural network” may perform classification or pattern recognition, but it does not understand in any human sense. Floridi and Nobre warn that the misuse of language can lead to category errors that affect ethics, policy, and education.
The epistemic and ethical stakes
Floridi and Nobre situate this issue within a broader epistemological frame. The metaphors we choose determine what we consider knowledge, explanation, and progress. If intelligence is framed purely as computation, then ethical questions about consciousness, autonomy, or responsibility are reduced to engineering challenges.
This framing has societal implications. When we anthropomorphize machines, saying that an AI “thinks,” “decides,” or “wants”, we may unintentionally absolve humans of responsibility for automated decisions. Conversely, when we “computerize” the human mind, we risk neglecting the embodied, social, and cultural dimensions of cognition. The language of technology thus feeds back into our understanding of ourselves.
From linguistic confusion to digital sustainability
Why should the digital sustainability community care about this? Because sustainability is not only about resources or carbon footprints, but also about cognitive and cultural ecosystems. A sustainable digital future requires conceptual clarity and epistemic responsibility. If we use the same words to describe brains and machines, we blur ethical boundaries, confuse accountability, and undermine trust.
In this sense, Floridi and Nobre’s work contributes to what we might call epistemic sustainability, i.e. the ability to sustain coherent, transparent, and responsible narratives about technology. Avoiding conceptual overreach is part of designing sustainable systems, since language guides how we build, regulate, and interact with digital infrastructures.
Towards a new linguistic ecology
The authors end with a call for balance. Rather than rejecting metaphors altogether, they suggest cultivating a critical awareness of their reach and limits. Interdisciplinary language can be productive, but only if we remember that metaphors are tools, not truths. Responsible innovation demands what Floridi calls “semantic discipline”, i.e. the care to use words precisely, and the humility to admit when analogies fail.
For AI and cognitive science alike, this means building bridges without erasing differences. Machines can be powerful aids in understanding cognition, but they do not replicate it. Humans can learn from computational metaphors, but they must also preserve the richness of meaning that cannot be coded.
Conclusion
The article Anthropomorphising Machines and Computerising Minds invites readers to pause and reflect on how language shapes our technological age. The crosswiring between AI and brain science has accelerated discovery, yet it also risks entangling thought itself. In a time when artificial intelligence increasingly defines what we value as “intelligent,” Floridi and Nobre remind us that the words we choose carry ethical weight.
For digital sustainability, the lesson is clear: sustaining our cognitive environment means keeping our language honest, our models transparent, and our metaphors in check.
Full Paper:
springer.com