Philip Steenkamp: New generation of bots clever, but not wise

Overnight, it seems, artificial intelligence chatbots came, saw and conquered. Suddenly they’re everywhere: writing cover letters for job applications, penning jokes and acting as your search engine valet-slash-co-pilot-slash-companion.

Automated chat isn’t new.

But this generation of AI chatbots, built around large language models, has made a big leap. The conversations feel authentic. The output can be indistinguishable from what a human might write — except for the speed. Ask ChatGPT to write a sonnet about catching butterflies on a May afternoon in Norway, and it cranks one out in perfect iambic pentameter in under 20 seconds.

It’s amazing — and unsettling. Predictions include a new age of enlightenment, leisure and prosperity on the one hand and grim dystopias where human connection loses its meaning and many of us lose our jobs on the other.

We’re left with a flurry of questions: Will bots allow dramatic advances in synthesizing knowledge, or fuel the spread of misinformation? Put powerful creative tools in everyone’s hands, or rob artists of their livelihoods? Level the playing field, or amplify bias and inequality?