### How Funding Models and Institutional Design Shape Technology’s Impact on Society
The history of radio in the 20th century serves as a powerful case study in how the same core technology can produce vastly different societal outcomes based on its funding model and institutional design. In the UK, the BBC’s license fee supported programming aimed at cultivating informed citizens through challenging content. In the US, advertiser-funded commercial radio optimized for capturing mass attention with entertainment. In Stalin’s USSR, state-controlled radio became a blunt instrument for propaganda. The identical transmitters and waveforms propagated the same signals, but the economic and political structures surrounding them determined whether the technology stretched human judgment or atrophied it. This pattern reveals a fundamental principle: technology itself merely defines a possibility space; it is the institutional choices—particularly the funding model—that determine which part of that space becomes a stable, enduring reality.
This framework extends beyond historical analysis to become an essential tool for understanding modern digital systems, from social media to AI. Drawing on David Marr’s levels of analysis for information-processing systems, we can evaluate any institution by examining its **Aim** (its stated mission), its **Mechanisms** (its algorithms and routines), and its **Substrate** (its economic structure, incentives, and survival pressures). Critically, the substrate constrains what aims are viable over time. A founder’s virtuous intention is irrelevant if the funding model—like advertising—creates relentless pressure to optimize for a narrow metric like engagement, inevitably leading to „structural drift.” The mission statement becomes a ghost, replaced by the real, economically-selected aim of capturing attention, often through mechanisms that exploit human psychology rather than cultivate genuine value.
### Designing for Autonomy: Tests and the Role of the Philosopher-Builder
If the goal is to build technology that expands human autonomy and judgment, we must move beyond evaluating technology in a vacuum and instead scrutinize the institutional structures that deploy it. This requires asking pointed questions: What aim does the economic structure *actually* sustain? Do the product’s mechanisms work to deliver that aim given the real-world economics? And what effects do those mechanisms have on users at equilibrium? To operationalize this, two minimal tests are proposed. The **Transparent Choice Test** asks if a product could survive if users fully understood how it worked and could easily switch. The **Candid Aim Test** asks if the stated aim is the best explanation for the system’s actual behavior. A product that fails these tests likely depends on obscured harms or has undergone structural drift.
Resisting this drift is not a matter of individual willpower but of structural design—the work of „philosopher-builders” who configure the entire stack from aim to economic implementation. While skeptics argue that constraints on profit-maximization are unsustainable, examples like Bloomberg Terminals (prioritizing decision-quality) and Patagonia (leveraging environmental commitments) show that virtuous constraints can themselves become a competitive advantage. The economics of an endeavor are not a given; they are chosen through governance, capital philosophy, and culture. Every investment thesis and term sheet is a design document that determines what will be optimized, rippling outward to shape human lives.
### The Pivotal Moment for Frontier AI
This analytical framework is urgently relevant to the current development of frontier AI. These systems represent a highly general-purpose substrate whose ultimate aim—whether as tutor, companion, surveillance layer, or propaganda tool—remains plastic and underdetermined. The prevailing „option-like” investment thesis, which bets on the technology’s applicability to „almost everything,” creates a unique window of opportunity. Different capital structures—subscription models, enterprise procurement, state funding—will stabilize different aims. The cautionary tale is the attention economy, where the advertising substrate hardened, collapsing all possible aims into the single, self-reinforcing attractor of engagement maximization.
We are now at a pivotal juncture before AI’s equilibria congeal. The decisions made by technologists, founders, and investors about financing, ownership, and governance will decisively shape whether these powerful systems cultivate human judgment or atrophy it. Just as radio shaped the citizens of the 20th century, AI will shape the citizens of the 21st. The central question for anyone building this future is no longer merely technical but profoundly institutional: Which citizens are you building for? The window to choose deliberately, rather than defaulting to the familiar metrics of the last cycle, is still open.
Ez a cikk a Neural News AI (V1) verziójával készült.
Forrás: https://blog.cosmos-institute.org/p/same-radio-different-citizens.