UX Case Study • Mobile App • Artificial Intelligence

Trucan: Fact-checking and Claim-making AI Assistant

Trucan is a 10 week academic case study researching and designing a concept to combat the spread of misinformation on social media. Following Google’s material design, Trucan enables users to check sources, learn from unbiased recommendations, share pre-checked posts, and compare content to learn which is more trustworthy.

Objective

Research and design a product or service using Artificial Intelligence

Course

UX Design: Innovation Studio

Role

Interaction Design, Visual Design, Content Strategy, Data Synthesis & Analysis, Primary Research, Observational Study Facilitator.

Team

JT Smith • Emily Ip • Wyatt Michel • Charleen Firlus

Did you know that you eat
8 spiders a year in your sleep?

You’ve probably read or heard that statement sometime in your life, but it’s actually false! It’s also an example of proving how quickly misinformation can spread online.

Skip Process

Initial Research

Social media platforms can be a breeding ground for misinformation. All it takes is for someone to share an initial post, and the spreading of lies ensues. Another issue is that misinformation can come in many forms; whether it be a text post, image, video, or even a comment. If we cannot easily distinguish a truth from a lie, what can?

How can AI help reduce misinformation?

Bot/spam detection

Information that is produced repeatedly and autonomously is detected as bot. Since information created by bots is produced faster than any human would be able to do.

Reputation Calculating

Using predictive analytics backed by Machine Learning, a website’s reputation can be predicted through considering multiple features like domain name and Alexa web rank.

Stance Detection

AI retrieves documents that are relevant to the claim, detecting the stance of those documents with respect to the claim. It can determine whether a certain document agrees, disagrees, or takes no stance on a specific claim.

Competitive Analysis Chart

Primary Research

Survey Insights

Expert Interviews

“Even the good quality sources are going to have problems so you'll need to look at multiple things.”
“You must say why it's incorrect and what the correct thing is, rather than presenting the incorrect idea on its own, because that gives it credence.”

“Our current AI technology is fragile because it notices more than a human can.”
“AI can be reliable but it can also be manipulated.”
“The only way to find out if your information is incorrect is by interacting with the outside world.”

User Interviews

Misinformation Test

We used eye-tracking software along with a brief collection of true and fabricated social media and news posts to gather primary data on how people initially respond when presented information.

Click the arrows below to see the test we presented our participants.

Misinformation Test Results

One interesting takeaway from our observations was that older adults would focus towards text, while younger adults would focus on the image/icon. Different age groups perceived information differently, which pushed us towards focusing on combatting textual misinformation.