Practical Research in Human-centred AI
Principal lecturer: Prof Alan Blackwell
Additional lecturer: Dr Advait Sarkar
Taken by: Part II CST
Code: HCAI
Term: Michaelmas
Hours: 16 (8 x 2-hour sessions)
Format: In-person lectures
Class limit: max. 20 students
Prerequisites: Further Human–Computer Interaction
Moodle, timetable
Aims
This is an advanced course in human-computer interaction, with a specialist focus on intelligent user interfaces and interaction with machine-learning and artificial intelligence technologies. The format will be largely Practical, with students carrying out a mini-project involving empirical research investigation. These studies will investigate human interaction with some kind of model-based system for planning, decision-making, automation etc. Possible study formats might include: System evaluation, Field observation, Hypothesis testing experiment, Design intervention or Corpus analysis, following set examples from recent research publications. Project work will be formally evaluated through a report and presentation.
Lectures
(note that Lectures 2-7 also include one hour class discussion of practical work)
- Current research themes in human-centred AI
- Mixed initiative interaction
- Labelling as a fundamental problem
- Program synthesis (vs “agentic” AI)
- Generative AI and knowledge work
- Transparency and explainable AI
- Bias and fairness
- Student research presentations
Objectives
By the end of the course students should:
- be familiar with current state of the art in intelligent interactive systems
- understand the human factors that are most critical in the design of such systems
- be able to evaluate evidence for and against the utility of novel systems
- have experience of conducting user studies meeting the quality criteria of this field
- be able to write up and present user research in a professional manner
Class Size
This module can accommodate upto 20 Part II, Part III and MPhil students.
Recommended reading
Brad A. Myers and Richard McDaniel (2000). Demonstrational Interfaces: Sometimes You Need a Little Intelligence, Sometimes You Need a Lot.
Alan Blackwell (2024). Moral Codes: Designing alternatives to AI
Assessment - Part II Students
The format will be largely practical, with students carrying out an individual mini-project involving empirical research investigation.
Assignment 1: six incremental submissions which together contribute 20% to the final module mark.
Assignment 2: Final report - 80% of the final module mark