Organic Bots and Authenticity

For a few years now, human traffic has made less than 50% of total internet traffic. This is annoying for people and companies who run websites trying to attract authentically human traffic to translate into views, customers, and sales. We have useful tools to manage this for now, but those days are numbered. Experts at Google forewarn about a day called “the inversion”, when human internet users are falsely flagged for being bots more often than bots, and bots behave more convincingly as humans than humans do – as far as the bot-detecting algorithms are concerned.

Google, Facebook, and Youtube determine which clicks, shares, and views are human by sampling from past human behavior, and they train their algorithms to recognize which is human and non-human. Some bots are easily detected, for example, because they don’t use on-page elements as much as humans do. There’s an arms race, however, and some folks are designing bots that can access all website content and simulate human clicks.

Sometimes the bots fool the algorithms. They slip past undetected and their behavior is taught to the human-detecting bots as human behavior. The sample of “human data” is tainted, more bots pass as humans, and the effects compound on each other until authentic human web traffic does not pass the “human test”.

We organic humans aren’t making things any easier, either. Our online media behavior is reinforced by likes and engagement from our friends and followers. We modify our behavior continuously to maximize our rewards. Take the collateral damage that comes with incentivizing comments on posts by offering rewards, points, etc. You get some of the most bot-like comments by people who are only commenting to catch that coin. Websites like Cent have a pretty decent sorting procedure to overcome this, and I’m most impressed with the site for this. But sometimes those organic bots slip by and get rewards. I’ve taken to seeding comments directly on Cent to reward the most thoughtful. In general, what we post and share on most sites (even Facebook) becomes more calculated, more predictable, … more bot-like.

I bring this up as a follow up to a previous post, “To Be a Better Writer, Don’t Identify as a Writer”. While the advice is specific to writing, it describes in the abstract a kind of inauthentic posturing that we all do in some way or another. Existentialists call it acting in “bad faith”, Erving Goffman described it as being on a stage, and superficially acting out our identities in the presence of others. The clothes we wear are costumes, the products we buy are props and our interests reveal our character traits to the audience.

Once I understood this, I spent the year of 2017 playing a character I named Justus Shepherd. I kept a journal owned by Shepherd and I recorded my experiences as if I were really him. He processed my experiences much differently than Coty would have because Justus is a much different character than Coty is.

Justus is a disgraced social sciences professor who lost his job after using his data to “perpetuate dangerous stereotypes about sex and gender,” as described by his university's termination order. One of his subjects reported him after learning that the survey data she submitted was used in a report that made conclusions she disagreed with. She felt that his research misrepresented her, and she raised a social media campaign to expose him. His university fired him without an appeal process and issued the following public statement:

“We greatly regret having approved Dr. Shepherd’s study without carefully considering the impact his results would have on marginalized communities. Our university is committed to training socially conscious scientists and his research does not represent our values. Although our Institutional Review Board rigorously evaluates the studies of our researchers to protect the individual participants from damaging or unethical conduct, we recognize the need to ensure that members of marginalized communities are protected from harmful stereotypes. We will work harder to evaluate all proposed studies with new guidelines to make sure all future research findings align with our values.”

Dr. Shepherd was barred from ever doing research again – at least in a professional capacity. I spent the year fleshing out this character and went back to being Coty again sometime in 2018. I started doing my own independent research under the occasional supervision of Dr. Shepherd. We would exchange letters and he advised me to organize a study to find out “what is it that makes intellectuals so damn unbearable?”

I did three small studies under this advice. In one, I modelled the evolutions of biological theory and sociological theory decade by decade starting in 1890 to the present. From 1890 to 1929, the sociological literature was focused on concrete problems pertaining to social organizations and a fair bit of self-reflection on its methods. The literature primarily focused on problems of war and peace and social cohesion starting in the 1930s to 1959—which makes sense considering there were two world wars up to that point. From that point on, gender theory, critical theory, and identity theory started to reach critical mass in the field. Much of the sentences that the scientists in these fields wrote reminded my of a sentence of Noam Chomsky’s I’ve read: “Colorless green ideas sleep furiously.”

Chomsky’s sentence was a proper and understandable English sentence, as far as grammar is concerned. It’s completely untethered to reality as far as its content goes, however. A good example of this from my sample, from “Bringing Identity Theory into Environmental Sociology” published in Sociological Theory in 2003, goes, “We conceptualize the environment identity as a person identity.” (I’ve made sure that my sample contains only journals with high impact factors to minimize what experts within the field would call, “a total bullshit paper.”)

The problems the researchers were addressing were less concrete, less empirical, and increasingly analytical. The theories in themselves are internally consistent and logically rigorous, but they lack what Chomsky’s sentence lacks: they don't bare any consistent relationship to reality. They describe relationships between elements of the theory that don’t necessarily represent anything outside of the theory. They are recursive.

I find it interesting that most of the analytically pure, empirically-detached statements are primarily coming from the identity theory crowd. It reminds me of Kierkegaard’s definition of “the self” as “a relation which relates itself to its own self, or it is that in the relation that the relation relates itself to its own self; the self is not the relation but that the relation relates itself to its own self.”

I’ll talk more about the methodology and the data of this study, and the others in their own separate posts, but I’ll pause here and hear what you have to say. Do you think we are becoming more bot-like? Do you have any experiences recognizing your own inauthenticity, and what did you do about it?