Life Off-Label

Big tech is learning an expensive lesson right now, and as a health outcomes researcher, I find it amusing.

The trend I have recently read about is called “tokenmaxxing.” Some companies have been pushing employees to use as much AI as possible to the point that some built usage leaderboards or tied AI consumption to performance reviews. The results? Uber blew through its entire annual AI budget in four months. Amazon quietly retired its usage leaderboard. In some cases, the AI bill now exceeds the salary of the person using it.

Health economics has a name for this: moral hazard. When the person using a resource never sees the bill, consumption goes up. It’s why we have copays. You’ve essentially built fee-for-service for AI. Volume is rewarded, but outcomes are not.

Imagine running a research team this way. A medical writer gets rewarded for cranking out drafts no matter how much editing each one needs. A biostatistician is celebrated for running the most analyses even when half of them answer the wrong question. Nobody would fund that program. Not because the people aren’t working, but because the work has no direction and no outcomes to show for the blown budget.

Those who know me know I’m not anti-AI by any means. I use AI in my own work and it can genuinely make me more productive. However, I’m purposeful about it. I don’t use AI for the sake of using AI. I use it where it helps me be more efficient with my time and gets me closer to my project goals.

Big tech is discovering something healthcare has known for decades. Reward volume, and volume is what you get.