English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Drawing from the resources of psychoanalysis and critical media studies, inthis paper we develop an analysis of Large Language Models (LLMs) as automatedsubjects. We argue the intentional fictional projection of subjectivity ontoLLMs can yield an alternate frame through which AI behaviour, including itsproductions of bias and harm, can be analysed. First, we introduce languagemodels, discuss their significance and risks, and outline our case forinterpreting model design and outputs with support from psychoanalyticconcepts. We trace a brief history of language models, culminating with thereleases, in 2022, of systems that realise state-of-the-art natural languageprocessing performance. We engage with one such system, OpenAI's InstructGPT,as a case study, detailing the layers of its construction and conductingexploratory and semi-structured interviews with chatbots. These interviewsprobe the model's moral imperatives to be helpful, truthful and harmless bydesign. The model acts, we argue, as the condensation of often competing socialdesires, articulated through the internet and harvested into training data,which must then be regulated and repressed. This foundational structure canhowever be redirected via prompting, so that the model comes to identify with,and transfer, its commitments to the immediate human subject before it. Inturn, these automated productions of language can lead to the human subjectprojecting agency upon the model, effecting occasionally further forms ofcountertransference. We conclude that critical media methods and psychoanalytictheory together offer a productive frame for grasping the powerful newcapacities of AI-driven language systems.

Structured Like a Language Model: Analysing AI as an Automated Subject