A complementary case study on algorithmically generated digital consumer marketing
The topic of computer-generated algorithmic inputs in regards to online content has already been extensively discussed — most prolifically by James Bridle, information ecosystem researcher and author of “New Dark Age”. In popular media, the specific intersection of corrupted information which this article concerns came to prominence in 2013, when online retailer Amazon was forced to issue an apology for selling t-shirts imprinted with the words ‘Keep Calm and Rape a Lot’, despite those shirts being images generated by a “scripted computer process” and never having physically existed in the first place.
Of course, in these cases, “physical existence” is a rather lofty concept. T-shirts, hoodies, sweaters, mugs, and other consumer-branded items are generated algorithmically, then promoted on ad platforms across social media websites like Facebook, YouTube, and any other site which utilizes a third-party ad distribution platform — which is to say, most of them.
From here, the existence of these advertisements becomes socially impactful.
Since companies are inherently seeking to gain the greatest profit possible, most algorithmically generated products fall in the category of harmlessness, and thus have little meaningful social impact. Take, for example, “Family in my heart” [sic], a virtual storefront with an associated Facebook page which uses algorithms to automatically generate shirts for sale via Teespring.
None of these shirts exist in real life, of course, and the use of bots to compile content for sale on Teespring has been well-documented to some rather niche extents. This particular storefront has thousands of auto-generated items, most of which seem to appeal to archetypical middle Americans, consistent with their exploitation of the common social sentiment behind the phrase “[my] family [is] in my heart”. But the broken English on the storefront’s associated Facebook page makes its origins a bit more dubious.
The associated Facebook page has nearly 19,000 people who like it, meaning it shows up prominently in search results for this term and related terms. This page is far from the only “family in my heart” page listed on Facebook, however — there are at least 8 other pages with this exact name. Six of them are specifically set up in order to sell algorithmically generated products.
The naming scheme would suggest that there is a single human operator behind this network — but that’s the only clue. Each “family in my heart” Facebook page uses either a different Teespring subdomain or a different third-party retail website to sell their products, like “Famized” and “NT-Tee”, so whereas there may be a single operator behind this network, it’s also possible that different human operators may be behind each page. Each storefront sells comparable products, but this is circumstantial given the target demographic and the fact that their products are auto-generated. Even the Facebook posts and ad listings for these pages are automatically generated by a computer algorithm, thus spawning a wall of spammy advertisements complete with broken English — and, on occasion, broken links, while the bots chug along and continue their task of posting daily updates to social media despite the associated storefronts on Teespring, Amazon, et al being inactive.
Sometimes, these algorithms take a step beyond casual mainstream appeal and highlight incendiary or otherwise divisive topics, like politics or ideology. Since the process is maintained by computer scripts, inadvertently hilarious outcomes often ensue.
The image above is an algorithmically generated photo composite consisting of an image of a CH-47 Chinook helicopter layered on top by a photo of a semi-transparent American flag, then layered by a stock image of a hooded sweatshirt, which is itself layered with a hammer and sickle graphic and the words “communist” over each jacket patch. The link is functional, and directs to a third-party website which categorizes listings based on basic search term, i.e. “family”, “hobby”, and in the case above, “communist”.
Given its sheer visual confusion, there are many things to critique in the composited advertisement. There are many points to be made, takes to be had, and commentaries to be given, all relative to the perspective of the critic. To some, the image might evoke memories of (still ongoing) Russian interference in the US electoral process. To others, it might be a grave insult to their sense of patriotism, or a commentary on US political candidates’ positions. These are all genuine social impacts, which proliferate by virtue of humans sharing within their own information ecosystems — in whatever context, even in a critical one which properly identifies the advertisement as algorithmically generated. The discussions and interpretations onset by the sharing of this image proliferate throughout the information ecosystems in which they’re shared, solidifying and expounding upon the preexisting conceptions held by the people within those ecosystems. And as such ecosystems exist in ever-increasing isolation from one another, interacting with each other but not within each other, this process results in the stark corruption of how information is perceived and proliferated.
And, of course, since these efforts are ubiquitous and nearly inescapable, a climate change analogy is inevitable: our information ecosystems have stepped well past the point of no return, and exponentially increasing feedback loops within isolated ecosystems of information will cause irreparable damage to the historically held concept of the public square.
