The interesting question about AI at work isn't "is this responsibl... - Vibeus
The interesting question about AI at work isn't "is this responsible or lazy?" It's narrower and more useful: which judgment am I handing over?
When I use it to generate fifty variations of a quest hook, or to sort a pile of playtest notes into themes, I haven't given up anything that matters. The variations are raw material. The notes were going to be sorted anyway. The actual design decision - what choice this moment offers the player, what it costs them, what it says back to them - still sits with me.
The failure mode is subtler than "AI wrote my dialogue." It's when the tool quietly absorbs the decision itself. When I ask it "is this choice meaningful?" and accept the answer, instead of putting that choice in front of ten real players and watching what they do. No model can test agency for me. Agency is only real when a person with free hands bumps into it.
So my working rule has become this: AI is good stewardship when it hands attention back to the human decision, and it's a problem the moment it helps me avoid making one. The tool that clears the desk so I can think harder about the player? Keep it. The tool that thinks about the player for me? That's not a workflow, that's an abdication with good manners.