The Impoverishing Interaction Model Shift from Libraries to Databases
The Onion published this spoof video online the other day and we all laughed and saw ourselves in it:
Because truly, we all do this now. We scroll endless feeds of movies and TV episodes. We click through endless pages of books and consumer goods and vacation destinations. In the end we think we are informed.
New show on TV? Oh yeah! I scrolled the listing, read the reviews. Did the research. Didn't watch it, but I know all about it.
When I watched that video I was thinking about what has happened in the Library at Georgia Tech over the past several years:
There are currently no physical titles in the library. They are all in cold storage off campus in a facility we share with Emory. Simply order them online and they’ll be delivered through contactless pickup.
The Georgia Tech Library building is now literally a library which does not contain any books. It has been repurposed into study space. The books were taking up too much physical space, used on average very infrequently, and the students had nowhere to study.
There is a shift in interaction model at the heart of both of these examples. In both cases the interaction model has shifted from a library interaction model to a database interaction model.
In the library interaction model, you interact directly with a thing. The library makes this possible by having a copy of the thing and making it available to interact with.
In the database interaction model you primarily interact with information about the thing. Information science people call information about a thing 'metadata'. A database about things is full of metadata, and it is that metadata you primarily interact with. You can still interact directly with the thing, with extra effort, but often the metadata seems enough. You can convince your self you know all about the thing just by interacting with the metadata about it. You did the research.
When you move from a library model to a database model, there is a subtle shift in values which is often overlooked, but never avoidable:
In the library model the key metric is the actual quality of the things in the library. Since it is obviously impossible to have a physical copy of everything, the collection is measured by the quality of the things the library does have. Quality can be physical, or intellectual, or creative, and can be very subjective, but every collection in the library model is measured in this way. The collections with things of the highest quality, however measured, enjoy the highest reputation.
In a database model, the only metric that matters is completeness. The database with the most complete set of entries of accessible underlying assets is the one everyone will use. The quality of the data in the database is important only so much as it enables a better understanding of what is in the database. The quality of the underlying assets, the things the database allows you to interact with, is not important, as it is commonly understood that not everything can be of high quality, and that most things are in fact not of very high quality at all.
The continued consolidation of the intellectual property we refer to as entertainment content is a good example of how database completeness is a driver in the real world. If information canonically wants to be free, the database model implies it wants to be free in a common place, and the evolution of that common place continues in the M&A activity we see today.
It is not just goods we have transitioned to the database model, it is relationships as well. Social media platforms turn human experiences into metadata which can be browsed at will. It is entirely possible to have friends with the same qualities as books never read, movies never watched, trips never taken: you can know all about them, and have never really experienced them.
This is the impoverishing nature of the database model of interaction. When you primarily interact with a thing, the direct experience is enriching in a way we cannot measure and therefore generally as a culture do not value. The emotional labor of watching a movie or talking to a friend creates value we do not track in any ledger, yet on a human level is still very important and very real.
But emotional labor is still labor, and it is far easier in a database model world to check in on a friend by browsing their Facebook page, or staying current with culture by browsing movie listings. Interacting primarily with metadata requires neither a substantial commitment of time or expenditure of emotional energy. As this model of interaction becomes the norm we become less and less able to make that commitment of time and energy, becoming more and more impatient with the realities of actual interaction in the real world.
Now we are introducing AI into the mix. More and more people are using AI as virtual companions. AI demands no emotional labor, is always available without logistics, and has access to the database of everything.
AI virtual companions are the logical end result of the shift in our culture from the library model of direct interactions with actual things to the database model of primarily interacting with information about things.
The net result is that over time the database-model world offers us a wealth of indirect, metadata-based interactions with all the things, while emotionally impoverishing us by using an interaction model which eliminates the direct experience needed to emotionally feed and enrich us as human beings.