The work of translation after ChatGPT and Large Language Models

The seminar studies the increasingly ubiquitous role played by Large Language Models (LLMs) and tools such as Chat GPT, Dall-E, DeepL and crowdsourced Online Translation Services, by investigating their historical development, technical form, and their current impact on society. Despite the increasing interdependence between translation and technology, the advent of machine translation has been poorly understood from a philosophical and political perspective.

Paraphrasing Naoki Sakai (1996), we should therefore ask “What sort of social relation is machine translation in the first place?” Translation cannot be seen simply as an operation of equivalency or transfer, but rather must be considered in a discussion about the production of social relations in which it “assumes a vital historical role akin to that played by labor in the constitution of the social” (Solomon & Sakai 2022). To understand machine translation—and the work of translation—is therefore to provoke a discussion about automation in general and the automation of knowledge and mental labour in particular.

The utopian dream of universal understanding across the globe – surprisingly enough – has been nurtured by politics of bordering and colonial expansion. This is true for the 1949 Weaver memorandum which opened the field of computer translation and was distributed at the dawn of the Cold War.¹ The leechlike Babel Fish from the Hitchhiker’s Guide to the Galaxy reminds us that linguistic barriers might be better off left unbroken in spite of cultural understanding if we don’t want to shed more blood.² Large Language Models like ChatGPT are set to transform the experience of border crossing: such is the case with college entrance exams, proof of language proficiency,³ and visa applications,⁴ all the while endangering linguistic diversity.

A credit certificate (Schein) is possible with credit in Media Philosophy. The course is addressed to the students of all departments of HfG and universities in the region. It is also open to students, researchers and guests from other institutions and “the world as classroom” (bell hooks) through the tools of the digital space.

Instructors:
Prof. Dr. Matteo Pasquinelli
Paolo Caffoni

In collaboration with:
George MacBeth
Arif Kornweitz

Course start: 26 April 2023
Course end: 19 July 2023
Wednesday bi-weekly, 15:00 – 18:00
Language: English and more

Consultation hours on appointment
Enrolment via Moodle or email
pcaffoni@hfg-karlsruhe.de

26 April 2023 15:00–18:00
10 May 2023 15:00–18:00
24 May 2023 15:00–18:00
7 June 2023 15:00–18:00
21 June 2023 15:00–18:00
5 July 2023 15:00–18:00
19 July 2023 15:00–18:00

The work of translation after ChatGPT and Large Language Models

The Hitchhiker's Guide to the Galaxy, 1981 BBC.