Searching for Bias

You are here

Last week, Larry Snyder (ISE professor and I-DISC Director) challenged a group of first year students to think about bias when they search the internet. In this hands-on workshop, Snyder helped the students explore some fundamental questions about internet searches: Are search engines objective catalogs of the contents of the web? Do they simply gather information about web pages and list them for us in neutral ways? Or do they replicate, amplify, or introduce bias when presenting us with search results?

The workshop was offered through Lehigh’s “5x10” series for first-year students and reflects I-DISC’s commitment to promoting equity and justice in data and computing. The topic was inspired by the work of Dr. Safiya Noble on algorithmic bias in search engines. The students at the 5x10 workshop did a simple experiment: Do a Google image search for “professor,” “nurse,” or “basketball player". Who do you see or not see? Do the images you see reflect real populations? What does this tell us about bias in search results? The group also discussed how search engines use artificial intelligence (AI) “under the hood” to identify images, including the use of machine learning (ML) to label images to “train” AI algorithms. The workshop concluded with a discussion about whether AI contributes to or mitigates bias.

In total, 37 first year students representing Lehigh University’s four main colleges attended this event.

5x10 is a programming series managed through The Office of First Year Experiences (OFYE) and open to all Lehigh students with a specific focus on first-year students and their connection to Lehigh. All first-year students are expected to attend five programs during the first ten weeks of the semester in at least three of the Five Foundations of Student Success: Collaborative Connections, Creative Curiosity, Identity Development, Inclusive Leadership, or Professional Growth & Success) in bLUeprint.