Data Creation and Collection for Artificial Intelligence via Crowdsourcing

Artificial Intelligence systems are increasingly powerful, yet their performance, fairness, and transparency often suffer due to limitations in the training data. When data is imbalanced, biased, or doesn’t reflect real-world variability, even the most advanced models can fail in subtle but critical ways.

This course introduces crowdsourcing as a powerful method to improve AI systems by involving people directly in the data creation and evaluation process. By leveraging human intelligence at scale, you’ll learn how to gather high-quality, representative, and meaningful data, essential for building robust machine learning models and trustworthy AI systems.

You’ll explore task design, quality control, and the use of active learning to enhance data collection, as well as ethical considerations when using human input in AI systems.

What you'll learn:

• How crowdsourcing can be used to generate and enrich training data

• How cognitive biases and human factors affect data quality

• The role of active learning in managing and optimizing human input

• How to design tasks for crowdsourcing with built-in quality control mechanisms

• How to evaluate machine learning systems with humans in the loop

Note: This course is hosted on an external platform (TU Delft via edX). You will be redirected to their website to access the content.

The ILIAD Academy is not affiliated with TU Delft or edX, and any accounts, progress tracking, or certifications will be managed solely through their platform.

Click here to access the course ➜