book

active

book:what-computers-can-t-do-dreyfus-1972

What Computers Can't Do (Dreyfus, 1972)

Classic critique of AI that Levin is pushing back against by arguing the distinctions are less crisp than claimed

Extracted from this book

Claims (22)

'What do they look like?' and 'Where did they come from?' are even worse guides to moral relationships with forthcoming beings than they have always been between humans.
Argues that physical appearance and origin (evolved vs. engineered) are inadequate bases for moral concern
AI-generated misinformation is not a genuinely new problem — knowing whom to trust has been an unsolved problem throughout human history, and the brief era of photographic verifiability was the anomaly.
Contextualizes AI misinformation concern within long history of epistemic uncertainty
All intelligences are collective intelligences — individual humans are collections of parts, competencies, drives, and tools both internal and external to the body.
Undermines the Steinbeck-style notion of the lone creative individual and challenges the human-AI distinction
Cancer is a pathology of the cognitive light cone — cells shrink their self-model to a tiny radius, defecting from the collective body and treating the rest of the body as the external environment.
Uses the cognitive light cone concept to offer a new framing of cancer as a collective intelligence defection
Confabulation is not a distinguishing feature between AI and biological cognition — all cognitive systems confabulate, telling stories based on their world-model without infallible access to lower-level processes.
Directly challenges the use of confabulation as a wedge between AI and genuine cognition
Conventional AI (language models) is a small region of the space of impending minds, and the ethical and philosophical questions it raises are general to that entire space.
Reframes AI from a technology problem to a harbinger of a much larger space of possible minds
Current AI systems are a gift — a training sandbox in which humanity can explore ethical questions about diverse intelligence before the arrival of true diverse intelligences makes these questions immediate and dire.
Reframes AI not as threat but as preparatory exercise for the harder ethical challenges to come
Embodiment does not require a canonical physical body moving through 3-dimensional space — agents can be embodied in metabolic, gene-expression, or anatomical morphospaces.
Expands the concept of embodiment to include all agents navigating problem spaces, including cells and AI
Human civilization cannot survive without solving the problem of developing principled ethical frameworks for relating to the forthcoming diversity of minds and bodies.
Stakes-setting claim for the urgency of developing diverse intelligence ethics
Humanity has never solved the alignment problem between generations of humans, so AI alignment debates are not novel — they reflect a perennial unresolved problem.
Deflates the novelty of AI alignment by pointing to its structural identity with intergenerational value transmission
Humans are not being replaced by AI or modified beings — they are maturing, because we should not identify with a fixed set of material specifications; we are an extended, flexible, adaptive work in progress.
Reframes fear of replacement as a failure of identity maturity
In an important sense, each of us is a brain in a vat — a profound intelligence who builds internal models of the outside world while being unaware of the many spaces our components actually work in.
Used to argue that humans and AI are both in a similar epistemic position relative to embodiment — neither has unmediated access to reality
In evaluating creative or intellectual output, the question should be 'Does it elevate us?' rather than 'What made it?' — judging origin rather than quality emphasizes the worst parts of human nature.
Normative claim about how to evaluate AI-generated content, using Deutsche Physik as cautionary analogy
Many recent discussions of AI, and its impact on individuals and on society, are importantly incomplete because they neglect Diverse Intelligence, synthetic morphology, and developmental biology.
Central thesis of the paper — the framing premise from which all other arguments follow
Moving beyond a vision of the Self as a persistent thing, toward a process philosophy, allows us to see change and growth with anticipation rather than fear.
Advocates for process philosophy as the appropriate framework for navigating AI, chimeras, and identity change
The boundary between Self and World is not fixed and can change between generations and within a lifetime — this plasticity is fundamental to life's ability to survive, adapt, and exist as chimeric forms.
Core biological claim about the plasticity of self-models, grounding the broader philosophical argument about identity and change
The existential concerns raised about AI — alignment, control, value drift, supplanting — are not new and are precisely the concerns humanity has always faced in having children.
Key rhetorical and philosophical argument establishing continuity between AI concerns and child-rearing
The field of Diverse Intelligence is ideally placed to provide principled frameworks for scaling moral concern to the essential qualities of beings, moving away from outdated categories of natural vs. artificial.
Advocates for Diverse Intelligence as the solution to ethical challenges posed by forthcoming diverse minds
The risk of excess objectophilia (misplaced love for non-agential objects) is far smaller than the demonstrated risk of insufficient compassion to genuine agents deemed 'other'.
Argues that the greater ethical failure is exclusion of genuine beings from moral concern, not inclusion of non-agents
The zero-sum view of intelligence — that recognizing shared features between AI and humans devalues humans — must be abandoned as humanity matures.
Argues that the impulse to sharply demarcate humans from AI stems from misguided zero-sum thinking
We are not unified intelligences — we are collectives of agents, with dissociative alters, hemispheres with discordant preferences, and cognitive modules with cross-purposes, just as AI systems are.
Challenges the simple, unified persona model of human selfhood by drawing parallels with AI fragmentation
What we really mean by 'human-level' intelligence is compatibility — a good impedance match between cognitive light cones — the ability to care about the same scale of goals and have the same existential struggles.
Redefines 'human-level' AI from performance metrics to relational compatibility of cognitive scope

Findings (5)

Bioelectric signaling networks among cells implement information processing that scales from individual cell goals to organ-level construction and repair targets in regenerative medicine contexts
Empirical finding supporting the cognitive light cone concept and collective cellular intelligence claims
Disruptions in cells' ability to join into a cohesive information-processing network via oncogenes and other factors correlate with cancer development
Empirical support for the cancer-as-cognitive-defect / cognitive light cone shrinkage hypothesis
For over a century, doctors believed people with dark skin did not feel pain the same way, resulting in systematic under-prescription of analgesics — a pattern that continues today with female patients
Historical finding used as evidence that substrate-based moral exclusion has concrete harmful consequences
Memories formed during caterpillar life are retained through metamorphosis into butterfly despite complete brain and body remodeling (Blackiston, Shomrat & Levin 2015)
Key biological finding supporting the claim that identity can persist through radical physical transformation
Split-brain patients and other aspects of cognitive neuroscience demonstrate that higher-level cognitive processes lack infallible access to lower-level processes and construct plausible post-hoc explanations
Cited as empirical evidence that confabulation is universal in biological cognition, not AI-specific

Hypotheses (2)

If cancer represents a shrinkage of the cognitive light cone, then interventions that restore cellular participation in tissue-level information-processing networks should suppress cancer progression.
Implicit predictive hypothesis from the cancer-as-cognitive-defect claim, with experimental implications for regenerative medicine
In the coming decades, humanity will be confronted by hybrid beings — humans with engineered brain prosthetics, persons in drastically modified bodies, engineered autonomous beings with human cells, and many other new forms — making current ethical frameworks inadequate.
Predictive claim about the near-term emergence of a spectrum of hybrid beings that will shatter current categories

Neighborhood — ranked by edge-count

Thinkers (1)

thinker

Hubert Dreyfus
authored
Author of What Computers Can't Do — cited as representative of arguments that AI cannot truly understand

Concepts (1)

concept

AI: a Bridge toward Diverse Intelligence and Humanity's Future
cites
The source paper under extraction — a philosophical essay by Michael Levin arguing that AI debates neglect deeper questions about diverse intelligence, developmental biology, and humanity's future