A Duck Test for High-fidelity Anthropomorphic Services

Abstract

In recent years there has been a significant increase in the number of anthropomorphic (human-like) services and interfaces with a high level of fidelity and realism. Renderings of human characters in photos and videos, chatbots, voice assistants and online influencers are examples of contents of digital mediums that are designed to reflect human looks, voices, conversations, movements, or communication styles. The implication for design is that service designers must now learn how to work with realistic simulations of humans in different aspects of their work. In contrast to cartoons or other non-realistic anthropomorphic representations, high-fidelity human-like characters used in consumer-facing services presents unique ethical challenges. It calls for upgraded frameworks for understanding the limitations and affordances of the novel materiality, which properly highlights the risks and opportunities with technology that passes for human. In this conceptual paper, I introduce a Duck Test for assessing whether a simulation of a human can pass for human or not. The duck analogy was originally minted by James Whitcombe Riley, who once wrote that ‘When I see a bird that walks like a duck and swims like a duck and quacks like a duck, I call that bird a duck.”. (Riley 2017) The Duck Test complements the Turing Test and the purpose is to ascertain whether a subject is believable enough to pass for the real thing.

Presenters

Gustav Borgefalk
PhD Candidate, School of Design, Royal College of Art

Details

Presentation Type

Paper Presentation in a Themed Session

Theme

Designed Objects

KEYWORDS

Turing Test, Anthropomorphic Technologies, Persuasive Technologies, Service Design, Interface Design