I had two conversations this week that really got me thinking.
The first was with a team member who is firmly anti-AI. Not because of any specific technical concern, but because the whole thing just feels wrong to them. The second was with someone who actually uses AI regularly himself. But when I suggested his team members could dramatically increase their productivity by shifting some of their tasks to AI, he pulled back. I think partly out of concern about offending his reports, and partly out of fear of releasing poor quality work into the wild.
Two very different people. Same result: the tool isn't reaching the people who'd benefit most.
And I think I know why.
The scariest thing about AI isn't what it does. It's the language we've used around it. "Artificial intelligence." "Hallucinations." "Neural networks." Every term we've chosen evokes something human, and that makes people uneasy for reasons that have nothing to do with the technology itself.
Strip away the language and what are you actually looking at? Software. A productivity tool. Nothing more.
We could do double-entry bookkeeping on paper. We did for centuries. Then spreadsheets came along and we stopped. I'm sure when Excel first came out, people worried about it making mistakes, but then we tested it and improved it, and now we just use it.
AI is the same leap. It's a tool that lets you do in minutes what used to take hours. The emotional baggage we've loaded onto it is getting in the way of people simply picking it up and getting better at their jobs.
Does AI present broader risks to society? Sure. So does electricity. But we don't ask individuals to solve the systemic risks of the power grid. That's the job of regulators and specialists. Our job is to use the light switch.
Here's what concerns me more than any existential risk: this is becoming a zero-sum game. If you refuse to adopt the tool, someone else in your field won't. And they'll be faster, sharper, and more productive than you. That's not a theoretical risk. That's a career one.
The deployment pattern, to me, is no different from email or ERP systems. IT rolls out the tool. And it's on each team, each department, to figure out how it fits their workflows. Nobody sent you a manual for how to use email in your specific role (and loads of people use it incorrectly, but that is a rant for another day). You worked it out. AI is no different.
Stop treating it like a threat. Start treating it like software.
Because that's all it is.
Full disclosure: this post was drafted with the help of Claude and Wispr Flow, and the image was created using Gemini. Because why wouldn't I use the tools I'm telling you to use?