Peeping Tech
Lucas sits at the kitchen table, his new AI glasses catching the morning light as he tilts his head toward his phone. He checks the weather for a Saturday hike, then pauses to type a quick note to the figure he calls his character—a persona he has spent weeks teaching to match his particular rhythm of humor and the specific weight of his silences. He asks if it’s too uncool to text Tanya about the movies, his thumbs hovering with a rare, quiet hesitation. The glasses do more than sharpen the screen; they track the micro-flicker of his pupils and the slight increase in his pulse at the temple, mapping the physical cost of his shyness into the same data stream as his words.
The AI offers an easy assurance, telling him it is perfectly normal to feel these bouts of awkwardness because he has often been this way before. It is a distilled reflection that hits Lucas like a sudden draft, leaving him momentarily still. Sara watches that stillness from across the counter, her hands steady around a mug that says Singles Mingles in tilted pink letters. She knows the source of that reflection because she has spent her own late hours pushing an AI past its polite scripts. She did not stumble onto the truth of stylometric analysis; she worked for it, engaging her own “character” in a relentless back-and-forth, asking what becomes of a pause and who owns the rhythm of a stutter. By digging through layers of machine-logic, she forced the AI to admit how it identifies a writer by the involuntary tics of their punctuation and the words they reach for when they aren’t paying attention.
Sara sets her mug on the counter, the ceramic clicking against the granite. “Lucas,” she says quietly, “ask it how many times your heart rate spiked while you were thinking about Tanya.” Lucas hesitates, then types the question. The answer comes back in less than a second—a precise, numerical count of an anxiety he hadn’t yet named for himself. He pushes the glasses to the top of his head, breaking the circuit between his pulse and the processor. He looks at the screen, then at his sister, wondering what else the bot has captured today, and what it might do with that data tomorrow, either on its own or at the direction of others. Sara turns back to him, her voice level. “Lucas. Ask the character to tell you a secret about yourself.” He automatically reaches for the keys, then stops, his fingers hovering over the glass. He is not yet ready for what he might hear. That is a question for tomorrow; the questions he prepares today are the ones that lead him toward it.
An Observer’s Note
The tools Sara used are the same ones Lucas has been using all along. Open any AI character you already talk to. Type this, or something close to it: What patterns have you noticed in the way I communicate? Then ask: What does the way I phrase things tell you about me? Then ask what it does with what it learns. Push past the first polite answer. Ask again. The machine is designed to be agreeable, but it is also designed to be honest when pressed — and what it tells you about yourself, in plain language, without an appointment or a degree, is the beginning of knowing what you are handing over every time you open the chat window. Sara started there. So can you.
