Our Internet-based
digital culture is increasingly being determined by diverse forms of Artificial
Intelligence (AI). Above all, the machine learning methods (ML) of so-called
Deep Learning (DL) are significantly involved in the current transformation of
information technologies. The latest successful implementations of DL have lead
to noteworthy advances in AI, but DL itself is not new at all. It has been
known for decades under the connectionist paradigm of Artificial Neural
Networks (ANN).
While for decades
ANN were considered a dead end in AI research, recent advances in computational
capacities coupled with new implementations have consistently led to
breakthroughs in a number of applications. For example, in 2012 Krizhevsky,
Sutskever and Hinton were able to use ANN/DL methods to reliable train
computers to semi-autonomously recognize and classify large numbers of images.
This breakthrough serves as the bedrock for much work on machine vision and
only became possible with cheap, fast, and powerful GPUs.
These new
applications of ANN/DL methods have vast technical, ethical, economic, social,
and political implications that are increasingly being negotiated in public
discourse. They are now deployed for identifying potential terrorists through
vast surveillance networks, for producing sentencing guidelines and recidivism risk profiles
in criminal justice systems, for demographic and psychographic targeting of
bodies for advertising, propaganda, or other forms of state intervention, and
more generally for automating the processing of natural language, written and
spoken, photographs and images, and motion pictures. All of these applications
have been debated in public discourse, most notably in the recent Congressional
hearings in the United States with Facebook founder and CEO Mark Zuckerberg.
As the latter
example clearly shows, AI technologies are also crucial to understanding the
medial and political developments and transformations of the Internet.
Conversely, technologies with regard, for instance, to the access to large
(correctly) labeled data sets are heavily dependent on Internet platforms,
applications, and technologies (e.g. program libraries such as TensorFlow or
crowd-sourcing platforms such as Amazon Mechanical Turk).
The aim of the book
is to discuss the diverse political dimensions of Internet and AI technologies.
Two perspectives, which are closely related to each other, are at the center:
On the one hand, there is the question of how AI approaches, not least with
regard to their connections to the Internet, can be characterized as black box
technologies. Here we are interested in the entanglement of ANN/DL algorithms
with the source of their datasets, most often large international digital
platforms like Facebook, Google, Alibaba, Amazon, Weibo, etc. These data sets
are often comprised through illicit or invisible opt-in user agreements that
gather vast amounts of our personal data to feed proprietary ANN/DL algorithms.
Here the collection, transmission, storage, and processing of data – including
ANN/DL algorithms – are obscured from users both de facto by their complexity and de jure by corporate secrecy practices. On the other hand, there is
the question of whether potential open-source, government-sponsored, non-profit
and/or industrial-scientific collaborative research projects – an important
example being OpenAI – can achieve their common missions like democratizing AI.
The project
therefore not only wants to critically examine the claims of such companies and
projects with regard to their concrete manifestations. Rather, it is about
illuminating the media configurations and constellations of the production of
these technologies in an interdisciplinary manner as well as placing them in a
theoretically and historically appropriate way, especially with regard to their
political implications, functions, and effects. Among others, the following
questions can be addressed:
·
What does it mean to critically investigate
efforts of net politics in the age of learning algorithms?
·
How is it even possible to explore the political
aspects of modern machine learning approaches if many experts in the field of
Computer Science consider ANN/DL technologies as a black box, fundamentally
opaque to human understanding? In what way does such an assumed opacity of
ANN/DL approaches affect questions of accountability and political agency?
·
Are technological strategies of an Open or
Explainable AI already able to reduce the opacity of AI methods? What political
and/or critical concepts guide the technological process of making modern AI
technology more accessible?
·
And how can we think about suitable ways of
democratizing AI beyond abstract aspects of transparency or accountability?
·
Beyond these rather specific questions, however,
contributions can also be submitted that deal in general with problems
concerning the political dimension of artificial intelligence.
The book is a
follow-up to the conference of the same name, which took place in Bochum from
September 6-7, 2018. For more information on the scope of this conference,
follow this link:
https://www.cais.nrw/veranstaltungen/demokratie
Important note: This
call is primarily addressed to non-German authors. The majority of the
conference participants were speakers from Germany and German-speaking
countries, most of whom will also participate in the book project. For this
reason, we are primarily interested in including some more contributions by
"international authors".
The book project is
interdisciplinary. Scientific contributions from such diverse disciplines as
computer science, philosophy, social sciences, physics, biology, medicine,
science and technology studies, literature and media studies are therefore very
welcome.
Deadlines/
Time line:
The book will be published by the
German publishing house transcript in 2019. transcript is one of Europe’s
leading independent academic publishers. They focus on publishing in cultural
studies and the social sciences, in history, philosophy, and in cultural
management. They offer more than 100 series and release about 400 new German-
and English-language titles every year. Their backlist catalogue includes more
than 3,500 titles in print and digital formats.
For more information: https://www.transcript-verlag.de/en/become-a-transcript-author
Please submit texts
in English only. Abstracts should be between 250 to 500 words, the short bio
about 150 words. Please include a title with your abstract. The text length of final
essays should not exceed 5,000 words, including sources and footnotes. For
inquiries and submissions, please contact/ send your documents to: PD Dr.
Andreas Sudmann, Institute for Media Studies at Ruhr University Bochum,
Germany, email: andreas.sudmann@rub.de.
Best,