The idea of the commons goes back further than technology, even property. In medieval Europe, the commons was the meadow where everyone’s cows grazed, the forest where families gathered firewood or the river where they fished. The commons was about survival, about living together. For generations, these places weren’t managed by distant officials or formal laws. They ran on ritual and trust. Rules like ‘don’t fish before the tide turns’, or ‘don’t cut branches thicker than your wrist’. The commons was less a thing and more a bond, a relationship between people, place and memory.
That relationship was broken in the seventeenth century. Across England and Europe, the enclosure movement turned shared land into private property. Fields were fenced and sold. Peasants lost their customary rights and were sometimes forced into cities or wage labour. What Karl Marx called “primitive accumulation” was for families a loss of home and belonging. And colonialism exported enclosure around the world—forests, pastures and rivers were seized, local traditions swept aside. The living commons became a story of loss and exclusion.
But the idea of the commons didn’t disappear. In the twentieth century, it returned in a new form: knowledge. When Garrett Hardin wrote about the “tragedy of the commons” in 1968, he argued that shared resources would always be overused because people act for themselves, not the group. His essay shaped policy for decades, justifying privatisation and strict control. Hardin’s argument overlooked centuries of successful commons governance.
Enter Elinor Ostrom, the political scientist who won the Nobel Prize in 2009. She travelled from Nepal to Switzerland to Kenya, showing that the “tragedy” was not inevitable. Communities can and do govern the commons, creating local rules, monitoring use and working together to solve problems. Commons, she proved, are not market failures—they’re living institutions built from the ground up, rooted in trust and reciprocity.
The digital age brought the commons back once again. In the 1980s, frustrated with companies locking up software, Richard Stallman founded the Free Software Foundation. He believed that software should be free, in price as well as in the freedom to use, modify and share. This resulted in Linux, open-source code and eventually projects like Wikipedia.
In the early 2000s, Lawrence Lessig and others developed Creative Commons licences, making it legal for anyone to share their work without giving up their rights. For a moment, it felt like the digital world could become a true commons: knowledge and culture shared by millions, kept alive by volunteers. Wikipedia remains a shining example—a digital commons, written and maintained by people everywhere, open to all.
But new enclosures have crept in. Data is the new land, fenced off by corporations that scrape it from the public internet, then lock it inside proprietary models. Companies like Meta, Google and OpenAI gather billions of words, images and voices—yet the results are hidden behind paywalls or subscriptions. What used to be free public knowledge becomes private insight. Talking about digital commons today means talking about this ongoing struggle: between open and closed, between care and control, between sharing and enclosure.
During the broader turmoil over digital culture and AI governance, open source has re-emerged as a critical idea—not merely a technical artefact but a political claim about who technology is for and how it should be governed, especially as proprietary platforms proliferate and subscription models dominate. Meta has released open-weight AI models, where the model’s knowledge and behaviour are known, so that developers can download and customise them. But critics note these do not fully meet classical open-source criteria because documentation and training-data transparency remain limited.
Chinese firms such as DeepSeek have released open-weight and permissively licenced models, with some offerings under the MIT licence and freely available weights that can be used and modified, enabling developers and researchers worldwide to experiment without exclusive dependence on a few proprietary providers. Today the open AI ecosystem includes many players, such as research collectives; independent developers sharing quantised models, adapters and fine-tuning recipes; and startup labs emphasising open development. Open-weight models have become globally diffused, redistributing some technological capacity across regions and actor types.
But while open source in AI offers real benefits, it also faces structural challenges that prevent it from fully fulfilling the promise of a commons. If open source was once a simple claim about shared code, AI exposes new complexities. AI systems depend on code, data and compute—code can be shared, data is harder to share transparently at scale and compute remains concentrated in a few hyperscale facilities. The resulting systems are often gated behind APIs, subscription layers or restrictive licences, so what was once a distributed knowledge commons can become monetised inference.
Within the open-weight ecosystem, transparency about training data and evaluation methodology is often limited, and some licences restrict commercial use or derivatives. In practice, the rhetoric of openness can outpace the reality of access and power. Open source no longer guarantees decentralisation; it has become one of the arenas in which geopolitical and corporate power is negotiated. Chinese models can enable local development without exclusive reliance on US cloud APIs, European research groups sustain open infrastructure with public support and independent developers adapt models at a pace that national strategies struggle to track. But this open layer remains unstable.
Without broader governance, transparency and infrastructure support, ‘open source’ risks becoming a symbolic label rather than a bulwark against enclosure. Worse, it could become a figurehead for what is otherwise a highly concentrated industry.
True progress will require stronger legal frameworks, enforceable transparency standards, data stewardship models and public compute investments so that shared power and shared capacity are not limited to rhetoric. The question is no longer whether open source exists; it is whether this ideal can survive scale. The most compelling answers are not coming from hyperscale laboratories, but from communities that treat language and data as collective inheritance rather than raw material…
If the commons is truly real, it has to be lived and felt in the hands and hopes of everyday people. Whether in bustling tech cities or quiet rural villages, you’ll find communities of artists or creators rolling up their sleeves to protect technology and culture from being fenced off. Their ways of doing can differ but they share a simple conviction: technology shouldn’t just serve a privileged few, and culture is too rich to be flattened into dull algorithms. These real-life stories—messy, inventive and sometimes fragile—remind us what’s at stake, and just how much can happen when people come together to build something for everyone.
In early 2025, Namibia took a bold step, launching one of Africa’s first national strategies for responsible AI. For a country of just 3 million, this was no small feat. Instead of copying Silicon Valley, with UNESCO’s help, Namibia planned to place diverse local voices and accountability front and centre. The message was that AI should serve ordinary Namibians, not just outside companies or global trends. The plan stood on three legs. First, establish the basics—better data access, digital government, stronger universities. Second, set the rule—new data laws, ethical guidelines, clear processes. Third, root everything in culture, acknowledging Namibia’s thirty-plus languages and diverse indigenous traditions as essential to how AI is designed and used.
The plan is being put into action. Researchers at the Namibia University of Science and Technology, led by Naftali Indongo and Heike Winschiers-Theophilus, are working side by side with indigenous communities to build AI that respects and reflects their lives. With the Ovahimba and San peoples, they’re co-creating digital tools that weave together local wisdom and practices with new technology. In one project, farmers helped create AI-smart hydroponics, a system that combines traditional ecological knowhow and machine learning, assisted by AI helpers named by the community—Ham-AI (‘Sun AI’) and Xaru-AI (‘Food AI’).
When Namibia’s national plan was launched, officials made it clear that they didn’t just want to buy technology from elsewhere; they wanted help in making it themselves. This vision is all about communities writing their own rules. Of course, there are still big hurdles: not enough computing power, reliance on donors and the loss of local talent to bigger tech hubs. Still, Namibia’s approach sends a powerful message: that ‘Responsible AI’ can come from the margins, too. Even a small nation, by drawing on its own culture, can conceptualise technology that’s about dignity, not just dependence.
Excerpted with permission from Westland Books.