Should You Use AI for a Task? Here’s a Simple Way to Decide

This article, originally published in The Guardian, discusses the use of artificial intelligence (AI) for various tasks and provides guidance on when it is appropriate to employ AI assistance. The author, a public policy professor at Harvard Kennedy School and the Munk School at the University of Toronto, argues that AI should be used for tasks that are similar to work, while reserving human effort for tasks that require creativity and critical thinking.

This analogy, drawn from AI researcher Daniel Meissler, suggests that using AI at work is akin to using assistive technology like a wagon or a forklift, while using AI at the gym—where the emphasis is on the process of performing the task—is analogous to weightlifting. The author provides examples to illustrate this distinction, noting that tasks such as writing policy memos, which involve critical thinking and skill development, do not benefit from AI assistance.

Instead, these tasks serve as a "gym," helping students develop essential skills that will be useful in their future careers. On the other hand, tasks that are purely mechanical or formulaic, such as writing an instruction manual, benefit from AI assistance, as these tasks are best suited to AI's strengths—accuracy and consistency.

The author also discusses the implications of AI's growing capabilities in the creative industries, such as writing, visual arts, and design. Historically, human writers, artists, and designers were relied upon for all types of tasks, regardless of whether they required work or creative effort. However, with AI's ability to handle work-type tasks efficiently, there is a shift occurring in the job market, leading to a potential decline in demand for human writers, artists, and designers.

The author argues that this shift requires a reevaluation of the roles these professionals play in society. Despite the author's clear guidance on when to use AI, students still resort to using it for tasks that are better suited for human effort. The author empathizes with students' reluctance to appear less competent in comparison to their peers who use AI, and acknowledges the challenge of maintaining discipline in the face of the convenience offered by AI.

The author concludes by emphasizing the importance of distinguishing between work and gym tasks in our lives, similar to making choices between using stairs or an elevator, or walking instead of calling an Uber. This distinction can help individuals make more informed decisions about when to leverage AI's capabilities and when to rely on their own human skills.