[NickyLloyd/Getty Images]

There’s a particular kind of silence that settles before history turns. Not a calm stillness, but the moment when something begins to shift beyond our awareness. We’re standing in that space now, between the internet we knew and the system it’s rapidly becoming. As 2025 closes, we humans are no longer the primary inhabitants of the digital world we built.

A new study by Graphite found that AI-generated articles surpassed those written by people sometime in late 2024. After three decades of the web as a global commons of human expression, most new content now comes from machines. The scale of the shift is astonishing. In two years, content creation flipped from a human-driven enterprise to one dominated by language models that write faster, cheaper, and more convincingly.

Another study by Imperva is even more jarring. It found that in 2024, automated systems overtook human users as the majority of all web traffic: 51% of online activity came from bots, of which nearly three-quarters was malicious. Behind the scenes, an army of automated systems now outnumber actual people, constantly probing our digital and cognitive defences.

Together, these studies reveal a simple truth: The internet is no longer a human-majority environment. For the first time, the stories and interactions shaping our online life are produced by entities that don’t think, feel, or understand consequences. This is not the dystopia some had warned about – it’s much stranger. The threat isn’t that machines turn on us, but that they quietly reshape the environment in which human meaning is made. Our thoughts and intentions are simply outnumbered online.

Part of the problem is scale. We write at human speed, clicking and reading at a pace our nervous systems evolved to manage. Machines don’t. They produce storms of automated traffic that wash over the digital landscape, altering it before we grasp what’s happening. Most machine-generated articles will never be read by a human, and most bots never seek a human audience. In this new landscape, human communication becomes a minority language in a world no longer designed for human comprehension.

The deeper issue is ecological. We think of the internet as a tool, but it has long been a habitat for our public reasoning; a place where we gather information, form opinions, build trust, and search for meaning. As that habitat fills with synthetic content and automated behaviour, the quality of human thought degrades because the environment itself becomes less human.

Some marginalised communities report that AI feels more understanding than the humans around them – not because machines have empathy but because parts of the internet became inhospitable long before bots were the majority. When a synthetic system feels safer than a human one, it’s not a victory for AI but a warning about the erosion of human connection in digital spaces. And yet the pattern is uneven since for some minorities, AI has amplified the biases they face.

Machines may outnumber us online, but they cannot replace our capacity for meaning or care. Bots cannot understand why a society needs truth or why some stories matter more than others. Only humans can rebuild the connective tissue that synthetic systems mimic but never feel. The question is not whether machines will overwhelm us, but whether we’ll allow our social relationships and the public sphere to further atrophy.

In 2026, we must centre human intention in our digital architecture and strengthen the bonds that make our societies resilient. That means demanding transparency in how platforms handle synthetic content; building public-interest algorithms that promote trustworthy information; regulating automated agents; and helping citizens navigate an information landscape where the first voice they hear may not be human. It also requires living more of our lives in places where we can encounter one another directly and rebuilding the habits of attention and care that the early internet once amplified.

The internet was never valuable because it was efficient. It mattered because it was a raucous, unpredictable global conversation where people could recognise one another, argue together, and occasionally discover common purpose. These are the forms of connection now at risk of being crowded out.

We can still defend that human core. But doing so means refusing to see ourselves as passive users and instead acting as stewards of the digital spaces we inhabit. A world where human voices are outnumbered need not be a world where they are diminished, as long as we choose connection over acceleration and remember that meaning comes from humans, not machines.

Chris Kremidas-Courtney is a senior visiting fellow at the European Policy Centre, associate fellow at the Geneva Centre for Security Policy, and author of ‘The Rest of Your Life: Five Stories of Your Future’.