Credit in an AI World

Machine learning is all the rage these days. Feeding datasets to algorithms and letting them work their magic is giving us self-driving cars, better cancer diagnoses, automatic tagging for your photos, and Go players better than the current world champion!

Machine learning has sparked lots of interesting discussions on things like machine morality and the eventual subjugation of humans to machine overlords. But something as simple as who gets credit for things in a machine learning world is already plenty interesting. Even simply how we talk about the fruits of machine learning will present interesting questions for copywriters.

Take Google Photo’s stylized photo capabilities. How does Google tell you the stylized photos it automatically creates from your photos? Fortunately for this exercise, the copy differs across different platforms!

iOS says the photo is auto created from your content which seems most accurate; notifications on Web credit the user for an automatic creation which doesn’t make much sense; and Android and Web take the easy way out and completely dodge the question! These variations are likely by design; I’d be interested to dig into the learnings the Google Photos team gleamed from these “who gets credit” copy differences.