This is the introduction for The Case for Responsible AI, a five-part series on how sustainability and impact professionals can use AI responsibly. Read the full series [here].
Recently, I was running late to pick up my 3-year-old daughter from preschool, and I had a full evening booked with family activities. It was all bad timing — I had an internal presentation first thing the next morning, and while I had a mess of notes to include, I hadn't organized them into anything coherent or presentable.
Before I ran out the door, I opened up Claude Cowork, shared my notes, the intended audience and outcome, and our internal slides template. In the morning, I logged in early to get ready for the presentation. I downloaded what Claude had created. I tweaked the language throughout, but the structure and design of the slides — a tedious task that can take me hours to get just right — looked great.
How do you calculate the value of four hours of time savings that allowed me to hang out with my family instead of work into the evening or night? That's not AI PR or some hypothetical future scenario, but a real, meaningful regular improvement for my family. And it's repeatable: I take a tedious task I don't enjoy doing and which loses nothing by being done by AI, and I put the machine to work while I clock out and head to the playground.
That experience is what shapes how I think about AI — not hype cycles or doomsday coverage, but the actual texture of using it day-to-day. And it's led me to a fairly simple principle: AI earns its place when it saves meaningful time.
Too often, AI is characterized as fundamentally good or bad. Inherently, it’s neither — it's a tool whose value is entirely dependent on how you use it (the dubious nature behind how it was built notwithstanding). A complex query uses roughly the energy equivalent of running a microwave for 10 to 30 seconds. If that query replaces three hours of manual research, the energy equation is clearly favorable. If it replaces writing one sentence I already know how to write, it obviously isn’t. The same logic applies at scale. Fifty low-value AI tasks per week that could each be done manually in a couple of minutes is a harder case to make than ten high-value tasks that each replace several hours of work.
Which raises the obvious next question: how do you actually make that call, consistently, in the middle of a busy workday?
The Case for Responsible AI goes deeper. The five-part Deep Dive explores the costs and benefits of AI use and a practical framework for deciding when to use it and how to do so with intention. We’ll also talk about where all of this might be headed. Start from the beginning [here].