Fair Robot Decisions in a User Study

ISEF Category: Robotics and Intelligent Machines

Subcategory: Other  ·  Difficulty: Advanced  ·  Setup: School Lab  ·  Time: 1 to 2 Months

The Hook

What if a robot had to choose which person to help first? That sounds simple until both people think they are the one who should win. Your project asks a real question in robotics, how do humans judge a machine that has to make tradeoffs? A small tabletop arm can become a testbed for fairness, trust, and rule design.

What Is It?

This project studies how people react when a robot must choose between two competing human requests. Think of it like a referee with a tiny arm. The robot does not just move parts. It also follows a rule set that decides whose request comes first, whose gets deferred, and how the robot explains its choice.

A constitutional-AI-style rule set means the robot follows a written set of principles instead of making random or hidden choices. You can compare different rule sets, such as first-come-first-served, equal split, or priority based on context. Then you ask users how fair, trustworthy, and acceptable each choice felt. The robot becomes a way to study human judgment, not just motion.

Why This Is a Good Topic

This is a strong science fair topic because you can test a clear variable, the decision rule, and measure a human response with surveys or ratings. The project connects to real problems in assistive robotics, service robots, and AI safety. You can learn experimental design, user-study methods, and basic statistics without needing a full robotics research lab. The topic also gives you room to make your own rule set and compare how people react to it.

Research Questions

• How does the robot's decision rule affect perceived fairness among users?
• What is the effect of explanation style on trust in the robot's choice?
• Does a first-come-first-served rule score higher than an equal-split rule for fairness?
• To what extent does prior ownership of the task change ratings of robot fairness?
• Which rule set produces the smallest difference in fairness ratings between the two users?
• How does the presence of a short written justification change acceptance of a difficult robot decision?

Basic Materials

• Tabletop robotic arm with simple programmable control.
• Laptop or desktop computer for robot control and survey display.
• Basic survey form on paper or Google Forms.
• Two participant chairs or marked standing spots.
• Printed task cards for the two competing user requests.
• Stopwatch or timer app.
• Consent form and participant instruction sheet.
• Spreadsheet software for recording ratings.
• Camera or phone tripod for documenting trials.

Advanced Materials

• Tabletop robotic arm with SDK or API access.
• Computer with Python and data logging tools.
• Survey platform with randomized question order.
• Microcontroller or relay interface for external task cues.
• Audio or text explanation module for robot responses.
• Post-study coding sheet for open-ended responses.
• Statistical software for ANOVA, mixed models, or nonparametric tests.
• Optional eye-tracking or video analysis setup for richer behavioral data.

Software & Tools

• Google Forms: Collects fairness, trust, and preference ratings after each robot decision.
• Excel: Organizes trial data and computes summary statistics.
• Python: Automates robot decision logic and saves trial-by-trial logs.
• ImageJ: Helps if you analyze video frames to measure response timing or gesture cues.
• JASP: Runs free statistical tests for comparing rule sets.

Experiment Steps

1. Define the exact decision conflict your robot will face, and make sure both users can understand it quickly.
2. Choose one rule set as your baseline, then decide what alternative rule or explanation you want to compare against it.
3. Build a trial structure that keeps the robot motion the same while only the decision rule changes.
4. Plan your fairness measure, including rating scales, open-ended questions, and any order you need to randomize.
5. Design controls that separate the effect of the rule from the effect of the robot's movement speed, wording, or visual appearance.
6. Decide how you will compare participant groups, summarize the ratings, and test whether differences are real.

Common Pitfalls

• Changing the robot's motion and the decision rule at the same time, which makes you unable to tell what caused the fairness rating.
• Using vague survey questions like "Was it fair?", which gives weak data and lots of interpretation noise.
• Letting one participant see the other's rating before answering, which can bias the whole user study.
• Giving the robot different explanations across trials without tracking that change, which confounds rule effects with language effects.
• Running too few repeat trials per condition, which makes one odd response look like a real pattern.

What Makes This Competitive

A stronger version of this project does more than ask people which rule they like. It tests whether the robot's explanation, conflict type, or timing changes fairness ratings in a measurable way. You can also compare user groups, such as students who get the resource first versus second, or use a simple statistical model instead of only averages. If you frame the study around a real design question in human-robot interaction, the project starts to look like research, not a demo.

Project Variations

• Test how fairness ratings change when the robot explains its choice in plain language versus a short rule statement.
• Compare competing demands that involve time, shared resources, or physical access to see which conflict type feels least fair.
• Swap the survey focus from fairness to trust or willingness to use the robot again, then see whether the pattern changes.

Learn More

• PubMed: Search for review articles on human-robot interaction, fairness, and trust in automation.
• NIH PubMed Central: Read full-text papers on user studies and decision-making in robotics.
• MIT OpenCourseWare: Look for free courses on human-computer interaction, experimental design, or robotics control.
• IEEE Xplore: Search abstracts for recent papers on robot ethics, social robots, and fairness perceptions, then use accessible full texts when available through school access.
• NASA TechPort: Browse robotics and autonomous systems projects to see how engineers describe decision constraints and system design.