Conversation

Notices

allison (aparrish@friend.camp)'s status on Saturday, 10-Jun-2023 07:29:47 JST allison
in reply to

especially love this quote where the authors are just like... "have you stopped for a second to consider why you're even doing this?!"
In conversation Saturday, 10-Jun-2023 07:29:47 JST from friend.camp permalink
Attachments
1. Traditionally, the goal of Artificial Intelligence research has been to create systems that would exhibit intelligence indistinguishable from human behaviour. Text-to-speech systems for instance, are evaluated on how natural and fluent the output sounds. Though intelligence and understanding should not be conflated with systems that exhibit humanlike behaviour (Bender and Koller, 2020), the human tendency to anthropomorphise convinces us of a machine’s apparent intelligence (Proudfoot, 2011). It is in part due to this longstanding goal of anthropomorphic systems that there only exists a small body of work that does not seek to anthropomorphise, despite the growing awareness of the harms that can arise from anthropomorphising systems.
  https://friend.camp/system/media_attachments/files/110/516/599/307/438/775/original/dde7714aa67f5752.png
- allison (aparrish@friend.camp)'s status on Saturday, 10-Jun-2023 07:29:48 JST allison
  
  this paper is so so good https://arxiv.org/abs/2305.09800 it outlines the properties of conversational AI systems that cause people to anthropomorphize them, and the reasons that makers of such systems lean into those properties (i.e. "engagement" = $$$), and why that's almost always a bad idea. the authors categorize it as a position paper but it's useful as a literature review too imo
  In conversation Saturday, 10-Jun-2023 07:29:48 JST permalink
  Attachments
  1. Mirages: On Anthropomorphism in Dialogue Systems
    
    Automated dialogue or conversational systems are anthropomorphised by developers and personified by users. While a degree of anthropomorphism is inevitable, conscious and unconscious design choices can guide users to personify them to varying degrees. Encouraging users to relate to automated systems as if they were human can lead to transparency and trust issues, and high risk scenarios caused by over-reliance on their outputs. As a result, natural language processing researchers have begun to investigate factors that induce personification and develop resources to mitigate such effects. However, these efforts are fragmented, and many aspects of anthropomorphism have yet to be considered. In this paper, we discuss the linguistic factors that contribute to the anthropomorphism of dialogue systems and the harms that can arise, arguing that it can reinforce stereotypes of gender roles and notions of acceptable language. We recommend that future efforts towards developing dialogue systems take particular care in their design, development, release, and description; and attend to the many linguistic cues that can elicit personification by users.
  minute repeated this.

Public

Conversation

Notices

Feeds