Should restricting AI access create new kinds of inequality?

The "We Are All Creators" paper (arXiv:2504.07936) reframes generative AI as "alternative intelligence and alternative creativity" — not mimicry of human cognition but a distinct form operating through mathematical pattern synthesis over collective human output. The central argument: "these AI models are fundamentally built upon the vast digital output of humanity." They are not independent inventors but "sophisticated processors of collective human creativity and knowledge — systems that synthesize and transform our shared digital heritage into new forms."

The copyright impasse. If models derive from collective, distributed input to which virtually everyone in the digital sphere has contributed, then individual ownership attribution becomes "practically impossible and conceptually fraught." How could one trace the influence of billions of inputs on a single generated token? How would one quantify each contribution — by volume, by impact, by originality? The sheer scale defies traditional models of individual authorship and reward. This is not a legal technicality but a conceptual impossibility.

The access imperative. The collective-knowledge framing produces a strong equity argument: "If these systems derive their capabilities from humanity's aggregated knowledge and creativity, then restricting access to them risks creating new forms of inequality." If the capability was collectively produced, restricting its use to those who can pay concentrates collective output into private gain.

The synergy prescription. Rather than prohibition, the paper advocates human-AI complementarity: AI excels at processing vast data, identifying patterns, generating variations rapidly. Humans excel at contextual understanding, ethical judgment, emotional intelligence, and "truly novel conceptual leaps." In creative fields, AI becomes "a tireless brainstorming partner, a generator of initial drafts" — democratizing creativity by "empowering individuals who lack traditional skills or resources to bring their ideas to life."

The paradox with skill formation. This collective-access argument sits in direct tension with the skill formation evidence. Since Does AI assistance actually harm the way developers learn?, democratic access to AI degrades the capacity to use it critically. The collective-knowledge argument and the cognitive debt evidence are in direct tension — broader access is both ethically imperative and epistemically dangerous. Since Does incremental AI replacement erode human influence over society?, the democratization of AI tools may accelerate the very disempowerment that restricting access would create inequality around.

The homogeneity risk. If AI output is drawn from collective knowledge, and since Do different AI models actually produce diverse outputs?, then democratized AI creativity could paradoxically narrow the diversity of creative output even as it broadens access to creative tools. The "alternative creativity" the paper celebrates may be alternative in mechanism but convergent in output.

Since Does training on AI-generated content permanently degrade model quality?, the proliferation of AI-generated content — the natural consequence of democratized access — may degrade the very collective knowledge base that makes AI valuable. The tail distributions that represent minority perspectives, unusual ideas, and niche cultural production disappear first.