AI Is Causing a Crisis of Agency - The Atlantic

Back in the web-traffic-obsessed days of 2018, at a time of dawning awareness of how easily audiences online could be manipulated and spoofed by bots, the writer Max Read argued that the internet had crossed a threshold known as “the Inversion.” Not only had bots proliferated across the internet; they had come to constitute it. In outnumbering humans, bots were also loosening everyone’s grasp on the very reality of online experience. “What’s gone from the internet, after all, isn’t ‘truth,’ but trust: the sense that the people and things we encounter are what they represent themselves to be,” Read wrote.<br>Today, “the Inversion” feels almost quaint. Autonomous AI agents roam the internet, answering emails, sending texts, and occasionally deleting the code repositories of entire companies. An endless library of chatbot-speak crowds out human-written words in every Google search. Bots are spinning up music and videos, conjuring bad poetry and prose, building websites, doing research, making transactions, writing plodding memos to your boss, solving geometry conjectures. Those AI outputs then ride the rails of an internet controlled by black-box algorithms. Computers talk to computers, producing information to train computers to sound more like humans or to better engage them. Humans type into the box, scroll, and wait.<br>AI is driving people insane in all kinds of ways. Its overwhelming speed and existential stakes have given rise to generalized malaise and hostility directed at the industry, to say nothing of actual cases of AI psychosis. But a lot of this is subtler—a deepening of the bewildering, corrosive feeling Read previously described. Culturally, the flood of slop, AI influencers, fake accounts, and AI tools is blurring the lines of an already post-truth age. A specific paranoia is in the air, an abiding concern about being manipulated, suckered, influenced. Stealth marketing campaigns, mercenary armies of bots, and paid clippers have led anyone or anything that appears dubious to be deemed a potential “psyop.” Cheap imitations of expressions of human creativity are easier than ever to fake. Sentiment, perhaps even popularity, is easier to manipulate. On top of all this is the push into agentic AI—a future we’re told will consist of an internet crammed with bots performing human tasks.<br>People who don’t feel empowered by all of this are unmoored. Across so many levels of culture, there’s a feeling of control slipping ever so slightly away. You, me, all of us, whether or not we enjoy or use these tools, are living through a crisis of agency. The agita and paranoia, even the excitement—over AI’s encroachment on work, education, art, and culture—is the by-product of a cultural and technological moment in which humans are sliding into a more passive role in many activities. One way to look at the generative-AI boom is as a massive societal experiment foisted on us by Silicon Valley, the animating question of which is: What is a human for?

When you start looking, you see the anxiety over agency everywhere. You see it in the reactions to the mass layoffs at places such as Meta in preparation for an AI transformation, in the coverage of venture-capital-funded, bulk-content-creation bot-army start-ups that proudly claim, “Never pay a human again.” You can sense it among the software developers who feel that their reliance on coding tools is eroding their skill set, in the executives who confess that they don’t know whether their AI spend is justifiable. Or when you read reports that medical journals are filling with made-up citations, or a study that suggests that chatbot use is degrading our thinking, or an announcement from Google that it will offer an alternative to its link-based search results: AI agents that can scan the web on your behalf and either bring back a canonical answer or send you personalized alerts.<br>The discomfort is playing out in real time. Last week, after the literary magazine Granta published the Commonwealth Short Story Prize–winning story “The Serpent in the Grove,” suspicious readers began to point out what they believed to be evidence of chatbot text in the story. Soon, two other Commonwealth Prize winners came under similar scrutiny, as people began running passages through AI detectors. (The Commonwealth Foundation first said in a statement that none of its prize winners had used AI, but then it issued a second statement suggesting that it is taking another look.) AI boosters celebrated the news as an example of the sophistication of current language models; skeptics viewed it as something of a slop tipping point. In a recent essay, the writer Sam Kriss described the experience of scrolling through websites right now: “The more I clicked around, the more I started to panic. There was nothing, no human voices anywhere, just thousands of versions of the same cheery demon. Am I alone out here? Something’s happened to the world; it’s all...