At first it felt like a simple chore: click images with a stop sign, type the warped words, reassemble a puzzle and tick “I am not a robot.” Those tiny, repetitive checks—millions of them—turned ordinary users into unpaid labor for big tech companies.
Luis von Ahn coined the phrase “games with a purpose” in 2004 to describe systems that channeled human intuition into tasks machines couldn’t yet do: labeling images, transcribing messy text, classifying content. His ESP Game matched strangers to produce image labels. Google adapted that idea for image labeling, and von Ahn later created reCAPTCHA, where solving CAPTCHAs also helped digitize words from scanned books. After selling reCAPTCHA to Google in 2009, he co-founded Duolingo in 2011, which used free language lessons to crowdsource translations and annotations—producing valuable language data that now trains AI and underpins commercial services.
What started as contributions to a shared digital commons was gradually consolidated and monetized. As Ulises Mejias argues, collective labor that might have stayed public was instead captured, packaged, and sold by a small number of firms.
That outcome clashes with the internet’s earlier, countercultural image. In the 1960s and 1970s many Bay Area thinkers imagined networked computing as liberating and communal. Stewart Brand’s The WELL, the slogan “information wants to be free,” and the rhetoric of figures like Steve Jobs and John Perry Barlow reflected a belief that digital networks could free people from old constraints.
But removing politics from technology proved naive. Historian Fred Turner shows that early digital communities often reproduced familiar hierarchies rather than escaping them. The utopian rhetoric gave way to monetization: search engines, recommendation algorithms, and massive data collection reshaped social platforms into sites of extraction rather than merely connection.
Platforms were built to gather and commodify user activity. Attention, interactions, and contributions became raw material sold back to advertisers and used to train proprietary systems. Mejias and Nick Couldry describe this as a “data grab,” likening it to a new, digital colonialism in which a common resource is harvested for the benefit of a tiny elite. Today’s AI models are the latest layer atop that extraction: aggregated human activity powers closed models and services controlled by corporations.
Three decades after promises of a free, independent cyberspace, the internet’s infrastructure, platforms, and revenue flows are concentrated in a handful of firms. For many critics, that feels like a betrayal: communities and shared knowledge built the web, then those gains were privatized and monetized.
There are signs of resistance. Communities oppose new data centers, gig workers organize for fairer conditions, and younger generations report frustration or outright disillusionment with online life. Surveys show many young people would prefer a world with less internet or see social media as harmful. That dissatisfaction is driving renewed demands for change.
Turner and others argue the answer is political, not merely technical. Instead of focusing only on advancing machine capabilities, society must decide how these systems serve the public good. The early counterculture dream overlooked governance, power, and who would control the tools and data. If the internet is to reclaim its promise, debates must move from tech alone to policy, collective governance, and democratic control over data and platforms.