Subproject English Studies

Coding Creativity in the 21st Century

Contact: Robin Auer

While powerful, human-like or even super-human AI (or what is usually called AGI) has long been a core theme of dystopian science fiction, it is only within the last decade that such intelligent machines have seemingly come within close reach. And while AI has recently contributed to both the subversion of democratic principles as well as the discovery of a new, powerful antibiotic, it seems that most applications of AI thus far are still rather narrow in scope as well as perceived success. More crucially even, those applications that are successful are usually the ones that are either focused on a specific goal, or work under human supervision and – more often than not – with additional human aid. One area where AI has traditionally struggled and failed, and which is now moving to the forefront of AI research is that of creativity, especially with respect to language and communication.

The frontier of what is possible in NLP (natural language processing) has been pushed forward ever more quickly in the last decade. As a result, even creative and artistic tasks are no longer deemed entirely out of reach for ‘intelligent’ machines. The moment seems close when science fiction will have come full circle in a twisted way, with ‘creative’ machine authors contemplating the human experience rather than the other way around.

While such scenarios may not be entirely impossible, it is nonetheless important to take a sceptical, more realistic view of the status quo of AI and its capabilities. Furthermore, AI and AI research present a return to certain schools of thought – particularly formal logic and behaviourism – that possess a rich tradition of critique, and which therefore need to be reviewed against the backdrop of recent technological developments and trends. The observation that current AI research is limited to these approaches will be considered as a systemic reason for the failure to come closer to proper AGI.

The project aims to achieve three separate yet highly interrelated objectives:

  • reflect and classify the current state and capabilities of ‘creative’ AI (with a particular focus on analysing and producing literary texts)
  • assess how AI-based technologies and methods (and the field of digital humanities in general) can contribute to literary and cultural studies (both in making available new answers as well as in offering fundamentally new approaches)
  • propose specific ways of improving upon current AI’s linguistic and literary competence and creativity, as well as ways of moving closer to AGI