A question that used to be unnecessary
Not long ago, a well-written text was, by itself, a sign of human effort. A sharp photograph implied a camera and someone holding it. A recorded voice implied real vocal cords, at some point, in front of a microphone.
That link between 'this exists' and 'someone made it' was so constant we never had to think about it. It was the ground that a large part of our trust in what we saw and read stood on.
That ground isn't as solid anymore. Not because it vanished overnight, but because, little by little, it stopped being an automatic certainty and became an open question instead.
What changed wasn't only the technology
It's easy to stay on the surface and talk only about tools: language models, image generators, voice cloning, synthetic video. All of that is real and keeps advancing fast. But what really changed is something quieter: the relationship between effort and result.
Producing a long text, an elaborate image, or a convincing video used to cost time, skill, or money. That cost worked, without anyone planning it that way, as a filter. That filter is weaker now. What used to take hours can be produced in seconds, and what used to require specialized skill can be produced from a simple written description.
This doesn't mean most generated content is false or harmful. Most of it isn't. It means that the mere existence of a text, an image, or a voice no longer tells us, on its own, whether a person was behind it in the sense we used to take for granted.
Why this isn't a purely technical problem
It's tempting to treat this as a detection problem: if technology can generate convincing content, then we need better technology to detect it. That race exists and will likely keep existing. But reducing the question to 'how do we detect what's artificial' leaves out something more important.
The deeper question isn't only technical, it's relational: what do we, as people, actually need to feel like we're talking to someone rather than simply receiving generated output? That need isn't solved only with a verification badge or a detection algorithm. It has to do with trust, with presence, with the sense that there's a life on the other side, not just a response.
That's why this isn't a problem engineering alone will close, even though it helps. It's also a problem of social design: what spaces we build, what rules we give them, and what we choose to prioritize inside them.
A new discomfort, still without a name
Many people describe a similar feeling when reading certain comments, reviews, or messages: a brief, almost automatic doubt about whether a person wrote it or something generated it. That doubt didn't exist a few years ago, at least not at this frequency or across this many different contexts at once.
We still don't have a comfortable name for that feeling. It isn't exactly paranoia, because most of the time there really is a person behind it. But it isn't the old certainty either. It's an uncomfortable middle ground we learned to live with without ever fully resolving it.
Why we're writing about this
We didn't write this article, or the ones that follow in this series, because we have a definitive solution. We have a partial, modest one, which we explain in the following articles and in our manifesto. We wrote it because we believe the question deserves to be taken seriously, calmly, without alarmism but also without minimizing it.
If this topic interests you, this is the first entry in a series of articles exploring different angles of the same question: what we lost, what we can do about it, and why, while there's no definitive technical answer yet, live conversation between real people remains, in its own way, a partial one.
