As expected of the Atlantic, the author misses the point.
The assertion of "Dead Internet" is not (necessarily) that 99% of everything on the net is made by one sinister literal AI run by the US government (though things might well converge to this point eventually). The point is that currently, less and less of the content is organic or real anymore.
"Bot" in modern Internet parlance is shorthand. It CAN mean a GPT-style program posting stuff on its own. It can also mean a human guiding said program, editing and cherry-picking before posting. It can also mean a bunch of humans in a room somewhere who have been paid to blend into various communities and spread propaganda and disinfo, for the sake of nudging opinion or just to muddy the waters. Sometimes in the employ of nation states, sometimes by a company just trying to sell a new pair of shoes. It can also mean the countless rogue individuals who make such posts just because they like to be disruptive.
It can also mean the so-called "algorithms" at work on every major platform now, a near infinite set of interlocking rules twisting and churning beneath the surface. What video will be recommended? What tweet will rise, and which will be buried? What information will be easy to come across, and what will be nearly impossible? This is a level of control and propaganda the tyrants of the previous era could only have dreamed of, and it is the reality of the modern Internet. Not simply to crudely censor or broadcast an idea, but to nudge the population this way or that in ways that are often almost undetectable.
The Internet used to be largely unfiltered, aside from basic trash collection. Now it is highly MANAGED, in ways that are never clear or honest, by agents that wear many masks. THIS is what is meant by "bots." The Atlantic article itself is a perfect example. What is the point of the article? To inform or empower the reader? Of course not. It's to MANAGE. To address the "Dead Internet" theory, a theory that's dangerous because it's true, and blunt it, to shift your opinion by "debunking"/"clarifying"/scoffing.
The Atlantic article was (probably) not written by an AI. It was (probably) written by a real human, working for other real humans. But why was the decision made, by those real people, to write and publish such an article? Because the "Dead Internet" theory is true: the Internet has become an enormous propaganda machine, exploited in turn by any number of powerful groups. The Atlantic is tied into those groups, and they want to keep the gravy train rolling. Maybe Kaitlyn Tiffany was specifically directed to write such an article. Maybe (more likely) she was hired because she was the sort of person who would write a dismissive article like this unbidden as soon as she encountered the "Dead Internet" theory (a "good cultural match for us"). Either way she's no longer an individual simply expressing HER views, but a tool for the small group of people that increasingly dominate the net to express THEIRS.
She's a bot.
And the Dead Internet is full of them.