The thing that worries me even more is that more and more diverse types of content will be susceptible to algorithmic amplification in more and more sophisticated ways. Right now it's still relatively easy to tell what is a bot and what isn't because it's hard to generate credible text and video (and many other kinds of) content, but that's rapidly changing. It's a game of cat and mouse clicks.

This pervades every single sphere of life. Soon, it will become more and more difficult to distinguish the truthfulness of digital representations. In Baudrillard's words, I suspect that big data is becoming hyperreal. It is not that data is misrepresenting reality, but that it is actively influencing it, changing it to fit its representations.

Furthermore (and somewhat orthogonal), distinguishing fake content is getting harder because you don't even necessarily need bots (technological capital). If I started a restaurant and I wanted to acquire the maximal amount of attention and I had no moral qualms, I could hire 10,000 MTurks (or MTurk-equivalents) to write reviews for me on Yelp, and it will drive a positive feedback loop of legitimate good reviews. (This is exactly what's happening with Amazon rankings and reviews.) Amplification isn't just algorithmic, it can be produced by the application of many different kinds of capital.