how do I talk about using AI without sounding like an AI evangelist?

A reader writes:

I lead a mission-driven nonprofit that does excellent work for communities that really need it. I’m proud of what we do, and we have a strong, consistent staff who do the work very well. We could, of course, use another dozen people or a million dollars, but we do as much as we can, as well as we can, and we pay our staff competitively for the field.

Recently, I’ve found a handful of AI tools that are surprisingly good at filling some of the massive gaps that inevitably exist in an underfunded nonprofit. For example, when a phishing attack started spreading through our organization, I used AI to help me write code that let us identify and clean up thousands of malicious emails. I’ve also used it to do complicated scheduling and data analysis that would ordinarily take me the better part of a week. Those aren’t jobs that would otherwise belong to someone on my staff; they’re things that would otherwise fall to me, badly and slowly. Using AI for them has freed me up to spend more of my time actually working with people.

And: I have significant reservations about AI. I’m concerned about its environmental impact, its labor implications, the way it’s being deployed by corporations, and any number of other things. I have not reached the conclusion that AI Is Good. I’ve reached the much narrower conclusion that, right now, some of these tools are making it possible for my chronically under-resourced nonprofit to do more good with the same number of people and hours in a day. I’ve made my peace with that for now, while reserving the right to decide later that the larger costs outweigh the benefits.

Here’s the weird part. I don’t want people to think that I suddenly learned Python and personally stopped a cyberattack, or that I’m staying up all night heroically blitzing through spreadsheets and schedules. When AI has materially helped me accomplish something, I think the honest thing is to say so.

But every time I say, “The robot did it” or otherwise explain that I used AI, I feel like I sound like an AI evangelist: Look at this amazing thing AI can do! And sometimes, annoyingly, the thing it did really is amazing. That still doesn’t mean I endorse the whole enterprise.

So: how do you talk openly and honestly about using a technology that is genuinely useful to you without turning every disclosure into an advertisement for that technology? Is there a way to communicate, “I’m telling you AI did this because I don’t want to take credit for work I didn’t do alone — not because I’ve decided AI is wonderful”?

Ironically, or appropriately, or both, I am asking you instead of asking the robot.

I think the key is to recognize that conversations about AI have to become more nuanced than just support/oppose. AI has already been incorporated into an enormous range of things, and refusing to use it is impractical for a lot of jobs. Not for all of them, obviously, and certainly not every use is defensible (personally, I’d put theft from writers and artists in the indefensible category, but not necessarily using AI for, say, coding assistance).

I’d say the questions we should ask are: is this particular application useful? Is it reliable enough for this specific task? What data are you giving it? What’s the risk if it makes a mistake and how are we guarding for that? Is the underlying training/data acquisition ethically acceptable? Is this meaningfully different from plagiarism?

Those answers will vary from one use to another.

It’s worth noting that in your case, you’re not using AI because you want to eliminate people; you’re actually using it because you don’t have enough people. You’re using it as a tool to absorb some tedious computational work so that you have more time to do the interpersonal work that actually requires you.

As for how to talk about it, you could look at it as a matter of attribution, not of endorsement — like the way you might mention that you used Excel to analyze something. You might say, for example, “I used an AI coding assistant to help me write a script that identified and removed thousands of malicious emails. I reviewed and adapted the code, but I don’t want to imply that I wrote the whole thing from scratch.”

And I think you can be honest and straightforward about your conflicting feelings! It’s reasonable to think some applications are hugely useful while remaining profoundly concerned about copyright, environmental costs, corporate power, and the ways it’s being built, and you can say explicitly, “I want to be up-front that I have significant concerns about AI’s environmental and other costs. But it was useful for this particular problem.” Acknowledging that it was useful for a specific task isn’t the same thing as saying that it’s an overall good.