Why are we trying to fix broken things with AI instead of just, you know… fixing them?

Before we try to enhance, optimize, or supercharge a process with AI, let’s ask ourselves: Is this process bullshit to begin with?

Bill Gates wrote this piece on AI’s revolutionary potential. In it, he declares, “I’ve been thinking a lot about how AI can not only boost productivity but also reduce some of the world’s worst inequities.”

Cool, I’m onboard with that level of optimism.

Digital designers and techno-optimists like me dream of AI freeing us from mundane tasks and democratizing knowledge and creativity. (Though some folks worry it will take away the things we enjoy doing.) We're also excited about AI as a creative partner, making us more productive. Seriously, I haven’t been this excited about a new technology since Larry Page and Sergey Brin illuminated the farthest corners of the internet with their first algorithmic search engine. Actually, that’s not true. Since I first laid eyes on the Gabocorp Flash site in 1997 (wow, that site looked a lot better in my mind’s eye than it does in the cold light of today).

But a particular passage in Bill’s article gives me pause. “I see several ways in which AIs will improve health care and the medical field … they’ll help health-care workers make the most of their time by taking care of certain tasks for them—things like filing insurance claims, dealing with paperwork, and drafting notes from a doctor’s visit. I expect that there will be a lot of innovation in this area.”

Using AI to file insurance claims? Let’s unpack that. What exactly are we asking the AI to do? Fill out medical and insurance forms so doctors and patients don’t have to. Nice, so doctors can focus on being doctors. Cool…

But… why are we filling out those insurance forms in the first place? As designers, before we enhance or supercharge a process/task/challenge, sometimes we need to ask ourselves:

Is this process/task/challenge bullshit to begin with?

“In 2022, there were 25.6 million nonelderly uninsured people” in the United States (ref).

Growing up in the US, my family was often uninsured or underinsured, leading to crippling medical bills and a constant battle to make ends meet. And during the times we were insured, the process for getting medical care was so convoluted and bureaucratic that by the time we got the care we needed, it was a miracle we were still alive to receive it. (And that’s without mentioning my parents’ language barrier.)

Sure, design an AI to fill out the forms.

Or, you know, fuck that, and fix the problem to begin with.

AI isn't the ‘enemy’. It’s our own impulse to look for cheap, easy wins while ignoring the cost to others that’s the real problem.

People sell AI as some magical fix for issues caused by broken systems—like the deeply unequal, profit-driven US healthcare system. Having AI fill out Byzantine health insurance forms? It’s a band-aid that hides the real wound, perpetuating injustice and inequality.

Look, I get it. Solving systemic problems is hard. Because they’re systemic. They’re entrenched, complex, and have stumped brilliant minds for decades, even centuries. But deploying AI as a quick fix means we don’t just fail to fix the root causes of dysfunction—sometimes, we make things worse. We legitimize the problems we should be solving.

Focusing on tech-based solutions alone distracts us from systemic flaws and hides injustice.

So before we run around looking for ways to fix things with AI, let’s recognize that manic rush to ‘fix’ things with it as a red flag, a clue something is fundamentally wrong. Remember the old saying, “Where there’s smoke, there’s fire”?? How about this: “Where there’s a knee-jerk reaction to fix shitty things with AI, there’s probably a deeper, shittier thing at the root of it.” It may not be as catchy, but it’s a useful counterpoint to unchecked enthusiasm.

Using AI to help people navigate Byzantine health insurance forms without addressing the root causes of dysfunction perpetuates systemic injustice and inequality, and further legitimizes a broken system.

Let’s be clear: AI isn't the ‘enemy’. It’s our own impulse to look for cheap, easy wins while ignoring the cost to others that’s the real problem. Same as it ever was.

Am I saying don’t use AI to make insurance forms easier today? No, that would be cruel to those already struggling. But we cannot stop there. Slap AI on it, call it a fix, and walk away? No. We have to dig deeper (or delve deeper, as ChatGPT might say). We have to acknowledge the root causes.

I could drop the mic here, but rants aren’t my style. I’m a designer, an optimist—I want solutions.

My initial instinct is to create a framework. A way to tell when AI is the right tool, and when it’s a sign we're dodging the harder work of fixing a broken system. If only there were some sort of matrix…

Bill Gates inspired this 3×3… and it might just change the way we think about AI.

My “AI DIVE Matrix” (for Depth, Impact, Visibility, Evaluation—catchy, right?) could bring clarity to those tough AI decisions. It might even reveal the policies and systems we desperately need to change. And yes, I have a two-step AI prompt that makes using the matrix easier.

This is part one of a two-part series. Part two is a nerdy read that offers up all the nerdy details that strategy and method nerds will totally nerd out on. It includes the AI prompt and access to a Google Sheets 3x3 tool, making it a valuable approach for designers and strategists looking to uncover which problems AI might be—and might not be—suited for.

Ready to learn about the AI DIVE Matrix? Continue on to Part 2.

Wil Arndt combines business strategy with creative vision as the Creative Director at OXD and ThoughtFarmer. With a career spanning over three decades, he has led projects for notable clients like Electronic Arts, Government of British Columbia, and Pokémon. Founder of the acclaimed digital firm Mod7, now part of OXD, Wil’s work has earned him hundreds of industry awards, including recognitions from SXSW, Applied Arts, The Art Directors Club, and The Webby Awards. His innovative approach has been showcased at the Venice Biennale and the Banff World Media Festival, pushing boundaries in digital storytelling and design. Beyond guiding teams through transformative change, Wil stands out as the only designer in the world (probably) with the rare experience of being paid to work as a pirate.