Here's an example that should hopefully be provocative. Let's say I work in ad-tech, and am hired to do an online advertising campaign for an online casino. Le

Here's an example that should hopefully be provocative. Let's say I work in ad-tech, and am hired to do an online advertising campaign for an online casino. Let's say I leverage some generative models as part of my campaign tuning, as well as some pretty sophisticated user tracking. At the end of the day, suppose my algorithms have done a good job of creating and targeting advertisements that result in conversions. Everybody wins, right? I make money, my client makes money, consumers learn about a product they are (evidently) interested in. Suppose further that the advertisement my algorithms generated does a good job of triggering cravings in gambling addicts (perhaps emulating the sensory experience of slot machines), and that my targeting found online communities focused on addiction. The problems with this should be obvious. At the core, you're assuming that humans are perfect little rational automata, when the reality is that we're sentient animals with a lot of maladaptive brain behavior that can be exploited in known ways. Advertisements are not simple informational notices, but precisely crafted information packets designed to elicit specific behavior.