The Evolving Role of Designers in an AI-driven World

Dan Saffer spent more than 20 years designing, at Adaptive Path, Smart Design, and four years at Twitter, working across robotics, wearables, and appliances. Two and a half years ago Carnegie Mellon asked if he wanted to teach, and he moved to Pittsburgh to join the Human Computer Interaction Institute. He describes the job as two-way translation: taking research out to industry, and bringing industry back to his students.

Ask him what's noise in AI and he starts with the hype around AGI, pointing to reports that companies aren't seeing the productivity gains they paid for. His colleagues at CMU trace the problem upstream, to project selection: teams keep picking work that's technically hard, financially unworkable, and not actually valuable to users. Add the expectation that a probabilistic system will be right every time and you get the disappointment the market is now living through.

He's also skeptical that prompting scales. Nobody's average office worker is writing seven pages of prompt engineering, and describing a design in text is absurd when your hand is already on the mouse. He argues for a continuum instead: AI working invisibly on personalization and navigation, open text for brainstorming, and direct manipulation for anything that needs a dial or a handle. Chat isn't the endpoint any more than the command line was. For designers entering the field, he sees three roles taking shape, the deeply strategic, the vertical specialist, and the AI-powered generalist, and thinks the industry is making a mistake by not hiring juniors, since they're the ones with time to actually learn these tools.