Machine learning has begun to "colonize" many different domains of modern life, including the classification of illnesses, the steering of cultural taste, optimizing search results, determining recidivism in the granting of parole, as well as scoring job applications and credit scores. What happens when we start to give over more and more of our social, political, and economic decision-making to machine intelligence? Relate your answer to the articles below: 1. Cramer, Florian. "What is 'Post-digital?" 2. Vincent, James. "The Invention of AI ‘Gaydar’ Could Be the Start of Something Much Worse." The Verge, September 21, 2017 3. Greenfield, Adam. "Machine Learning" in Radical Technologies:

The rise of ever more embedded digital systems has brought on what Cramer describes as the "post-digital," in which the "newness" of computing has become an historical fact and new "analog" experiences are unironically manufactured and distributed by digital systems. The nature of such experiences is, by the nature of digital systems and their interaction with users, impossible to predict, but (according to Greenfield) will likely be different than that intended by their designers.

Expert Answers

An illustration of the letter 'A' in a speech bubbles

The James article describes an example of a post-digital system: a machine learning algorithm that can identify sexual orientation based on facial recognition. The article points out that such systems are akin to the pseudo science of physiognomy, which attempted to connect intelligence to head shape and facial characteristics.

Cramer would argue that such a system is "post-digital" in that it repurposes or reimagines an old system in digital terms. For Cramer, post-digital does not refer to a time after computers, but instead a time when computing technology is so ubiquitous that its "otherness" has worn off, and it begins to adopt or consume earlier "analog" technologies. Cramer's example is somewhat less alarming: he mentions the insertion of "glitches" in music to replicate analog "grain" even while the means of distribution remains digital. But the AI-powered "gaydar" system from The Verge article is another case of an old idea being resurrected by modern tech.

Adam Greenfield cautions, in the conclusion to his book, that systems, once they come into use, cease to be under the control of their designers. His point is that once AI is widely distributed, its effects can no longer be predicted and will, most likely, have outcomes very different than what the designers intended. James touches on this in his gaydar piece, suggesting that such systems in the wrong hands could be used for mass persecution of gay people.

Taken together, these pieces suggest that the ubiquitous computing resources of the post-digital society will be deployed to recapture and augment experiences through digital means.

Approved by eNotes Editorial Team