Institut für Informatik | Sitemap | LMU-Portal
English
  • Startseite
  • Studieninteressierte
  • Studierende
    • Ersti Infopoint
    • Semesterplanung
    • Prüfungsleistungen und Studienordnung
    • Abschlussarbeiten
      • Bachelor
      • Master
      • Projekt
      • Freie Themen
      • Richtlinien
    • Benutzerstudien
    • Auslandsaufenthalte
    • Vorlagen
    • Glossar
    • FAQ
    • Ansprechpartner
  • Lehrveranstaltungen
  • Forschung
  • Personen
  • Kontakt
  • Besucher
  • Jobs
  • FAQ
  • Intern

Publikations-Information

Designing a Socially Aware AI Assistant

master thesis

Status open
Student NA
Advisor Philipp Thalhammer, Marco Gutfleisch
Professor Alexander Wiethoff

Task

Preliminary Abstract

Artificial intelligence (AI)- powered assistants have become ubiquitous across many contexts and are used both privately and professionally. While we expect other human beings to adapt their behavior to the social context of a situation (e.g, not revealing private information in the presence of others or adapting the loudness of their voice), AI assistants currently fail to meet the same standards. Prior work has explored how social qualities affect human-AI interactions, showing that higher contingency leads to greater perceived trust, sincerity, and understanding, regardless of whether the AI is labeled as expert or non-expert. Similarly, it has been shown that when a system is perceived as 'warmer' or 'more capable', users tend to exhibit more positive feelings towards it. Beyond improving the quality of human–AI interaction, context-aware behavior can also actively mitigate potential harm and strengthen privacy in AI-mediated interactions, for instance by adapting interaction modalities to the surrounding social context.

Goal

In this thesis, you will build an AI voice assistant in the form of a tangible artifact that can sense social context (i.e the number of people present in a room) and is able to switch input and output modalities accordingly. This interface will then be evaluated in a user study to find out wheter users perefer an intelligent system that automatically adapts to a situation or control the system themselves.

What you will do

  • Find existing literature on (social-) context aware AI systems
  • Design and implement a intellugent assistant in the form of a physical artifact
  • Design and conduct an evaluation of your artifact
  • Write a thesis documenting your process and its findings

What we expect

  • You have experience with hardware prototyping (incl. working with microcontrollers, 3D printing, etc.)
  • You have experience with software development, ideally working with AI enabled systems
  • Solid skills in English reading and writing
  • You can work and solve problems independently
  • You are creative

What you get

  • Two committed supervisors, weekly meetings, and hands-on advice
  • Being part of state-of-the-art research on AI-user-interaction
  • A masters thesis

Keywords

Tangible interfaces, AI Interaction
Nach oben
Impressum – Datenschutz – Kontakt  |  Letzte Änderung am 11.04.2020 von Changkun Ou (rev 35668)