Steps to take before Intelligent User Interfaces become real

1
Steps to take before Intelligent User Interfaces
become real

• by Kristina Höök

2
Why we need Intelligent User Interfaces?

• Information overflow problems
• Help using complex systems
• Real-time cognitive overflow problems

3
Problems

• better usability
• methods
• intelligence to improve interaction
• better tools
• -gt intelligent user interfaces aiding the
  computer-human interaction process

4
1) Usability principles

• Control transparency and predictability
• domain/ internal/embedding transparency
• Privacy and trust
• system takes tasks on from the user
• Treating systems as fellow beings
• shift users perception of responsibility
• agents providing right expectations

5
Methods

• Meet the users needs
• A proper analysis of users

6
Analysis phases in designing adaptive systems

• functional analysis
• data analysis
• task knowledge analysis
• user analysis
• environment analysis

7
(No Transcript)
8
2) Methods

• cognitive task analysis CTA
• by Roth and Woods, 1989
• unreliable user characteristics
• individual differences linked to the problem
• users knowledge keeps changing

9
3) Design practice

• Challenge
• Improvement of interaction with the user
• Adaptive features/general usability
• Evaluating systems difficult

10
4) Scalability

• Problem of AI systems have not tackled the right
  problems
• New artificial intelligence systems can learn
  and extend their coverage by themselves
• How to take into account new user interests,
  preferences
• -gt combinations of human and machine
  intelligence, authoring tools

11
Summary

• Challenges
• Usability
• Development methods
• Useful adaptions
• Maintainability
• Design of systems must meet many different user
  groups needs
• Solutions rendered should actually solve the
  problems they claim to solve