While the case you describe might have been an accident, the big problem with the marketing-surveillance state Internet is becoming is that someone somewhen may - or maybe already did - decide to run such ads on purpose. After all, what's a better market for diet pills than people suffering from bulimia or anorexia? What's a better market for anti-depressants than people with suicidal thoughts?

I don't get your last example - getting antidepressants to people with suicidal thoughts sounds great to me.

In any case, while I have no doubt that people have and will continue to make such decisions, I doubt they will become the norm - they're too prone to lawsuits, and too inefficient. Instead, Google et all will continue to develop algorithms that detect and exploit those correlations from raw data, so that no one will be able to tell why your profile was selected to be shown diet pills. Which is much more insidious, since it diffuses the moral responsibility.

It sounds great on its surface, but antidepressants have a common side effect of increased suicidal tendencies.

You might be interested in the 15th episode (of the 4th series) of Person of Interest, "Q & A" (Aired at February 17, 2015), in which depressed and addicted people were targeted with special adds to drive profit up. I won't spoil the whole episode, but it certainly made me think about our implicit, almost tacit trust in search engines, social media sites, and other services we don't fully understand (as users, because we don't have access to the source or data), but which we use daily. I can only imagine our dependence on these services to grow in the future, while our understanding of these complex systems does not grow at the same rate (or at all).

That was exactly the episode I was subtly referencing in my comment. BTW. after watching it, I thought that the way the data was collected in that episode is a writer's fiction. Imagine my surprise when I learned few days later that it's actually true, and it's what Apple and Microsoft do.