DESIGNING ASSISTANT TECHNOLOGY

A multi-pronged analysis of the ways in which artificial intelligence can be harnessed for good.

The technologies we think of collectively as AI are “growing at a breakneck pace,” writes Noessel; in his view, they’re all centered around five “universal assists” that are unlikely to change: perception, knowledge, planning, performance, and reflection. These assists will hold true across the whole spectrum of AI functionality, per the author. Noessel uses a straightforward analogy to differentiate between AI assistants and AI agents—think of the difference between a guard dog and a hunting dog. The hunting dog works in tandem with its owner along very specific prescribed lines, whereas the guard dog has a much broader remit and exercises a much larger degree of autonomy. The author concedes that these designations can blur, acknowledging that “sophisticated products bridge and transition between assistant, agentive, and even manual modes of use.” In all cases, he maintains, the most basic question is, “How can the product assist people?” Using charts, graphs, cartoon illustrations, and a recurring fictional character named Chuck, Noessel lays out a series of strategies for integrating all types of AI into preexisting human organizations and spells out the risks involved. Discussing such risks, the author puts forward his book’s strongest element: his informed optimism. He believes that AI can make people smarter, and that designers can play a key role in that process. “How amazing to think,” he writes, contemplating employers replacing workers, “an enterprise might choose instead to superpower its people.” This upbeat hopefulness carries the book; Noessel isn’t simply wishing away the dark side of AI assistants (agentic or otherwise); he’s convincingly pointing out the range of benefits that could result from taking a more humanistic path.

A smart, warmly encouraging look at the future of AI design.