Truthfully, I had to put 'Model Behavior' down a couple times because certain chapters felt like watching a car crash in slow motion. It’s less about big, dramatic ethical breaches and more about the tiny, insidious compromises that accumulate. The protagonist, a data modeler named Leo, gets assigned to tweak a predictive algorithm for a client’s hiring platform. The requested ‘tweak’ is to subtly downgrade resumes from certain zip codes. The book brilliantly dissects the rationalization process: it’s just a weighting variable, it’s what the client wants for ‘cultural fit,’ it’s not technically illegal. The challenge isn’t a mustache-twirling villain handing down an order; it’s the project manager hinting at a promotion if deliverables are met, and the quiet knowledge that someone else would do the job if you refuse.
What stuck with me was the portrayal of isolation. Leo can’t really discuss the ethical unease with colleagues without sounding naive or jeopardizing his position. The office culture frames everything in terms of efficiency and client satisfaction, making a moral objection seem like a personal failure to be a ‘team player.’ The book explores that hollow feeling when you justify something you know is wrong by telling yourself it’s just a job, just code. It ends without a clean victory, which feels depressingly real. He manages a small, subversive act of resistance that likely changes nothing systemic, and the ambiguity of that is the whole point.