Week 3 Synchronous Session Questions
Question One:
Generative audio restoration is starting to become more widespread and normalized. From adding voice to the trenches of World War One in the movie ‘They Shall Not Grow Old’ to allowing The Beatles to release one final song, the ethics of digital restoration, and in these cases resurrection, are everywhere. What do you see are the most concerning ethical issues for the use of artificial intelligence in archival recordings, and what might we do to mitigate those risks?
Question Two:
Disembodied, time-shifted voices surround us. From station announcements to gratitude for shopping upon checkout, voice is a vehicle through which we attempt to humanize our interaction with technology, especially transactional technology. It’s always fascinating to read the stories which put faces to those voices, but we shudder at the myth of the canned laughter in our sitcoms being the voices of the dead. With generative technology, voice simulation is even eliminating the need for human participation beyond training the initial model as with Siri and Alexa. As the cost barrier to entry comes down in generative audio and voice replication, and more generative audio enters our political messaging, can any material form of regulation be implemented to help the public understand the real from the synthetic?
Question Three:
In our case study of photography as an emergent technology, we’ve talked a lot about the dynamics of affordability and adoption. And how what we may perceive to be established technologies still support existing social hierarchies of wealth and cultural division. Willingness and ability to pay are key for scaling, as much as they were in 1920s Turkey as they are in modern-day Philadelphia. But in this ability also comes the privilege of being able to retain. Much of your work in archival audio concerns what remains. The artifacts which have survived over time. But even now much of what is created online has an incredibly short half-life, making it challenging to archive even the most robust of platforms. And while there are obviously efforts to preserve, they are still dependent upon co-operation from the original source. What do you think are the biggest challenges for preserving the audio of the web, and how might we approach solving them?
