Uloženo v:
Podrobná bibliografie
Hlavní autor: Antonopoulou, Caterina
Médium: Recurso digital
Jazyk:
Vydáno: Zenodo 2026
Témata:
On-line přístup:https://doi.org/10.5281/zenodo.19369375
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Obsah:
  • <p>Recent research demonstrates that AI systems are neither infallible nor objective. They embed biases, reinforce stereotypes and often reach erroneous or over-simplified conclusions. This paper aims at investigating AI’s errors and biases through the interactive media art installation Zackie PIA. Zackie introduces itself as a Personal Intelligent Assistant (PIA) designed to assist journalists to write better reports about the cases they cover, by retrieving and analysing all relevant data. However, Zackie uncritically combines information from online sources (news sites, official reports, social media, and more), according to a custom algorithm functioning as an automatic narrative generator. As a result, it composes fake and self-contradictory reports. The installation aims at highlighting how algorithmic bias embedded in intelligent devices affects the decisions made by these -normally considered as neutral- technologies. Additionally, the installation reflects on the ways that anthropomorphized Artificial Intelligence (AI) devices such as PIAs and chatbots reinforce gender stereotypes by reproducing normative assumptions of gender and femininity. To critically approach this issue Zackie PIA was named after a LGBTQ+ activist and artist, whose murder and the subsequent surge of fake news were used as a case study to ‘train’ the algorithm of the ‘intelligent’ device. This paper summarizes major findings from recent research on AI's errors and biases and discusses media art and design fiction projects addressing these issues. Furthermore, it presents the artwork Zackie PIA, describing the concept of the installation, its technical implementation, the artist's design decisions and key observations on viewers' interaction with the artwork.</p>