What Claude Can’t Generate

In an AI world, how do we know what is real?

K. Yoho Photography / Shutterstock

SOMETIMES I HAVE what they call “lucid dreams.” I’m aware that what I’m experiencing is a product of my mind and not tethered to outside realities (like the law of gravity). I can wake myself up—with much effort—by yelling in my dream, which actually activates my real vocal cords (to my husband’s annoyance). As in the films The Matrix and Inception, the million-dollar question arises: How do we know what is real and what is all in our heads?

With the skyrocketing use of generative AI in our everyday lives—from work reports to dinner-table Googling—the question takes on new layers. How can we tell when something is generated by Anthropic’s Claude or OpenAI’s ChatGPT? What are the telltale signs of the “uncanny valley” experience? And—more importantly—is there something distinct and irreplaceable about our embodied reality that AI cannot simulate?

Read the Full Article

To continue reading this article — and get full access to all our magazine content — subscribe now for as little as $5.95. Your subscription helps sustain our nonprofit journalism and allows us to pay authors for their terrific work! Thank you for your support.
Subscribe Now!