Building memory in the information age

May 18, 2018

Building memory in the information age

Julia Rijssenbeek
May 18, 2018

Building memory in the information age

Julia Rijssenbeek
May 18, 2018
Photo courtesy of Onur Buz. © Unsplash.

Building memory in the information age

May 18, 2018

For the first time in history, a total recall scenario has become reality. We store an exploding amount of data and every bit of it can be recovered. Technological advances like high-capacity storage make selecting or filtering information almost superfluous. However, our increasing reliance on online data storage begs the question what building an archive means and what the role of forgetting is.

Our observations

  • Philosopher Boris Groys sees memory as the base of innovation. He calls an archive a place where older productions are preserved and that allows us to compare the old with the new. Groys doubts whether the internet can function as an archive in its current form, because it lacks historical memory and is largely dominated by a few private parties.
  • The fact that humans still forget makes them superior to AI. Machine learning algorithms still need to learn how to distinguish between outdated and irrelevant information, and information that should be kept for future applications. AI is, for example, known for either over-fitting, the problem of storing excessive amounts of detailed information about past experiences, hindering its ability to generalize and forecast future events, or catastrophic forgetting, being unable to appropriately adjust to new information without abruptly forgetting what was learned before.
  • In the Black Mirror episode “The Entire History of You”, all characters have access to a memory implant that records everything they experience. The fact that they do not forget anything turns into a nightmare of either escaping to or being haunted by all past experiences; characters live in the past and under constant surveillance. The value of forgetting becomes painfully clear in this scenario where everything is recorded.
  • Flash memory and the rise of edge data centers are accelerating and decentralizing data storage. Flash memories are the key storage technology for mobile devices, and flash storage capacity is roughly doubling every 18 months. With “edge computing”, substantial computing and storage resources are placed at the edge of the internet network, so that data streams do not need to travel the internet backbone and can be processed and stored locally instead.

Connecting the dots

Over the course of history, humans have relied on using external memory to support their personal memory, from the first written words to the first photographs. Today, technological developments have given rise to forms of artificial memory that endure almost indefinitely. We are witnessing an explosion of data and the global demand for high-capacity storage is exploding in order to feed the trend in big data processing. In this age of information, the internet offers the promise of a collective memory, generally accessible to all. Nevertheless, according to philosopher Boris Groys, endlessly storing information about things is not the same as preserving the things themselves. In Über das Neue, he uses the concept of “the archive” to describe the collection of everything that we preserve as meaningful, the collection that structures our culture, like traditional art museums and libraries. Choosing to add an object or an experience from everyday life to the archive is a conscious decision to save things for the future. The advent of online culture challenged this traditional idea of the archives. Everyone can endlessly (and mindlessly) create, store and recollect information in the archives of the internet. Classic mediators and curators of the archives no longer seem to be needed. Furthermore, Groys observes how with the fragmentation of online archival space, we are entering a “new virtual Middle Age”, where individuals are engaged in a series of self-installations and navigate a number of heterogeneous spaces (e.g. personal websites). Moreover, our online spaces are subject to privatization. The internet today is mainly dominated by a few private tech giants. They valorize the online archives, while traditional archives, such as museums and libraries, that used to have an image of being there for the public, are now struggling financially. These developments have led us to a transition period, in which we try to maintain classical structures such as museums, libraries and universities, while the impact of online archives is growing and the relationship between the two is a vague and uncertain one.By contrast, in an age of big data and increased digital storage capacity, it seems it is no longer possible to forget something. However, memory is a duality of remembering and forgetting. Especially in times of an overload of information and information avoidance, mechanisms to separate the relevant from the irrelevant are crucial. A common explanation of forgetting is that memories help us understand the world, rather than merely remember it. In this way, we seem to retain memories that are useful, valuable and relevant, while we forget information of lower value. There is thus no intrinsic reason why remembering should be given precedence over forgetting. Simply deleting information is not enough. While humans are capable of “smart forgetting”, AI still has to be trained to know when to keep old information and when to discard irrelevant information.Exploding storage capacity and the possibility of total recall do not make building archives and forgetting redundant. On the contrary, both are needed more than ever to navigate a world of endless information.

Implications

  • Advances in placing data centers at the very edge of wireless networks point towards more intelligent ways of storing data. When data increasingly moves to the edge of networks, the less relevant data can be kept out of central online archives, creating a sort of selection mechanism.
  • Decentralized technologies can offer the promise of a collective memory accessible to all and challenge the prevailing online archive dominated by a few private tech companies.
  • As we have written before, digital data hermeneutics, the theory and methodology of interpretation, will become ever more important in the future to transform the explosion of digital data into meaningful information.
  • The problems AI has with forgetting make neuroscience and the understanding of how our brains decide what is worth remembering and forgetting important in creating better AI. On the other hand, neuroscience is working on human memory modification and trying to figure out how to cure dementia. According to the UN, dementia sufferers will triple by 2050. In the future, natural and artificial memory will move closer together in light of these developments.
About the author(s)
At FreedomLab, Julia Rijssenbeek focuses on our relationship to nature, sustainable and technological transitions in the food system, and the geopolitics of our global food sytems. She is currently working on her PhD in philosophy of technology at Wageningen University, investigating how synthetic biology might alter philosophical ideas about nature and the values we hold, as well as what a bio-based future may bring.
You may also like