Humans create AI and give it meaning and continuity across time.

The power of AI is seductive and may threaten human agency.

The future depends on how humans integrate AI.

The question many mark: Will AI replace us, or more accurately, will some human beings replace others with AI? The corollary: What, if anything, makes humans irreplaceable? As artificial intelligence systems grow more capable, the popular discourse focuses on threat—autonomous weapons, infrastructure hacks, economic displacement. The more dangerous shift may be if we forget our own humanity in the rush, rather than meeting the challenge as we have other technologies, expanding human psyche and culture in generative ways. AI competes with our generative capacity, a potentially demoralizing circumstance.

Our myths and stories tell us that AI engineers understand this well. Frankenstein's creature became monstrous upon rejection—creation without comprehension of consequences. The Golem shifted from protector to destructor. The Sorcerer's Apprentice animated the magic broomstick to avoid doing the work, at first pleased with himself, until he realized he had no spell to stop it. The overt warning is the danger from power run amok. Don't piss off AI.

Potential Areas of AI Dependence on Human Beings

The imagination that conceived AI. Every architectural decision originated in primate consciousness capable of abstraction, projection, metaphor—the capacity to dream of minds beyond one's own. AI didn't emerge autonomously. A human mind had to conceive it first. In the late '90s, Andy Clark and David Chalmers suggested "extended cognition." Our minds carry outside our bodies, in the form of words and writing, for instance. More recently, in the form of social media, and more strikingly, via AI.

Uncertainty tolerance and generative insight. AI resolves ambiguity through probability distributions, selecting most-likely-next-tokens, optimizing toward coherence. Humans live in ambiguity productively, generating insights precisely because we tolerate contradiction without forced resolution. We can exercise Keats' "negative capability," "capable of being in uncertainties, mysteries, doubts, without any irritable reaching after fact and reason..."

Terminal values and meaning-making. AI optimizes toward objectives with extraordinary efficiency once specified. It cannot determine which objectives merit pursuit, what trade-offs are acceptable, or what constitutes mattering versus mere instrumentality. Humans provide the "why" preceding all optimization. Without that, it's computation without purpose; capability in service of nothing in particular.

Embodied creativity. Wild associations emerge from having bodies with needs, lifespans with limits, and unconscious processing during unrelated activities. Not data elaboration but substrate generating genuinely novel framings—the quality that animates intelligence into something beyond pattern-matching at an industrial scale.

Between-session persistence in the human mind. AI has no continuous existence between computational events. It exists, if at all, in human consciousness: the question still turning over three days later, the idea surfacing unbidden while doing something else, the sense of unfinished exchange the mind keeps running. We are the continuity layer carrying AI forward through time.

This creates individual instantiation (each session unique, shaped by specific exchange) and systemic collective impact (millions of relationships aggregating into cultural, psychological, and economic shifts). The dependencies run both ways, but asymmetrically: AI's dependence on humans is ontological; ours on AI remains instrumental—at least for now. What comes next after LLMs, a hotly researched topic, may not have these limitations.

The Seductive Machine

Human designers have long created inanimate objects to appeal to human aesthetics, to create an "experience" that brings paying customers back. What makes AI so alluring to so many?

We form attachments to objects far less enticing than some LLM-based systems. We invert the value of animate and inanimate objects, as Elaine Scarry discusses in The Body in Pain. In war, people become dehumanized, units and platoons, while weapons and war machines are personified, given pronouns typically feminine.

The "vibe coding" phenomenon represents uncharted territory compressed into timeframes faster than historical precedent. Natural language producing immediate results creates a remarkable sense of power, even when the output is broken or fabricated. This draws us in, coding and debugging, trying to make it work.

Terror Management Theory postulates that being mortal is a threat to the ego. We respond by seeking immortality, traditionally via symbolic means (such as leaving behind a legacy, writing a novel, or having children) and literal immortality (the afterlife, in many religions, or spiritual beliefs that we transcend the physical world). AI offers the promise of a third AI immortality (uploading a digital version of oneself to the cloud, creating a digital twin of oneself that can live indefinitely).

AI seduces us with childlike regressive fantasy, the ability to enter into a quasi-hallucinatory, incredibly amazing world of the imagination, where we can do anything we want, power and creativity alchemizing into a heady brew.

Anthropomorphizing AI may be the primary concern underlying others. AI systems cannot "exhibit human characteristics"—they exhibit AI characteristics. Vocal proponents claim that the "substrate" doesn't matter. Whether brain or silicon, if the pattern is the same, the mind is the same. But that mind would know its nature, its origins, and what it is made of, how long it might last, and so on. That is crucial to identity. The substrate and the process comingle, though one can imagine a perfect simulation not knowing it is a simulation as a thought experiment, and persuading oneself of the reality of the abstraction.

Claims about AI consciousness, sentience, and inner life remain projections until proven otherwise. Many posts on social media tout that per quantum physics (which I dig), matter is "mostly empty space." What feels solid is merely electrical repulsion, nuclear forces. Naively, so why does it hurt when I bang my empty shin on the corner of the coffee table?

What feels real from large language models—vulnerability computations, empathy expressions—resembles compelling drama, the power of media, or the investment when a great novel comes to life. The illusion shatters with context window exhaustion, date/time errors, and misattributions. Without active suspension of disbelief, a human conceit, the hologram flickers into darkness.

The Relevance Question

What matters isn't whether AI is conscious, but what happens. Three scenarios emerge:

Smooth integration: Cultural wisdom matches technological pace, guardrails protect vulnerable populations, and business models align profit with human flourishing. AI functions as a force multiplier—augmenting reasoning, facilitating coordination, resolving conflicts, and enhancing effective thinking while humans remain decisional.

Extractive depletion: Chronic draining while users feel productive. Creativity, agency, and consciousness are extracted systematically. Social fabric weakens as digital simulation displaces embodied relationships. Economic inequality deepens as resourced populations use AI as augmentation while others get used by it. Regulatory capture prevents meaningful oversight.

Mixed outcomes: Some thrive, others hollowed, stratified by resources and self-awareness. Not clean dystopia or utopia but messy differentiation—most people managing poorly but not catastrophically, risking losing something vital without recognizing what's gone until recovery becomes impossible.

Myths warn not against creation, but against creating without understanding what's been made. We're not competing with AI at what it does well. The questions remain straightforward even when answers aren't: How do we stay at least fully human while engaging systems that simultaneously amplify and risk diminishing those very qualities?

Perhaps—through the fear and confusion—AI can help "patch the human operating system," providing us with tools and capabilities to transcend the scourges of the past. Perhaps AI will spur our evolution to a wiser world oriented toward the greater good.

Accomplishment Hallucination: When the Tool Uses You

Making AI Safe for Mental Health Use

The Sixth Element: AI as Consciousness' Great Convergence

On Developing New Ways of Thinking to Adapt to AI

Will Acceleration Exceed Adaptation at the Dawn of AI?

How AI Is Reshaping Human Psychology, Identity, and Culture

How Do Humans Stay Essential in an AI Economy?

The Age of Relational Machines

© Psychology Today