TRUE OR FALSE!
FIGHTING MISINFORMATION TOGETHER
ROLE
Product Designer
TOOLS & SKILLS
Figma
User Research
Collaboration
OBJECTIVE
Integrate an AI-powered solution to address the problem of the spread of misinformation among journalists and knowledge-seeking communities.
TEAM
Me
Siyu Liu
Presentation deck→
GUIDING QUESTION
How can we help journalists combat the spread misinformation?
OUR APPROACH
This was a problem posed at the Hacks/Hackers (H/H) hackathon. The event brought together a large group of engineers, journalists, and designers to come up with solutions together. This presented a perfect opportunity for my team and I to conduct some live user research.
RESEARCH FINDINGS
All of the journalists present at the event were either pursuing a Master's degree in journalism or were already working professionals. This provided a rich pool of context to pull from.
INSIGHTS
We gained a few key insights. Most journalists only look to scholarly sources, and they are also understandably wary of trusting AI when it comes to accuracy.
CONNECTIONS
We found that affinity bias played a decently large role in how/why journalists connect with each other. This takeaway would continue to inform our design process.
THE PERSONA
A persona that we decided to focus on was that of a young journalist seeking accurate information in a sea of uncertainty. This person is still in school and looking to get the facts, expand their network, and contribute in the fight against misinformation.
INITIAL INVESTIGATIONS
We knew that we would integrate a chatbot into our product, but we wanted it to do more than just answer questions. It had to help the user get accurate information, prove where/who it was coming from, and offer opportunities to foster connections within the journalism community.
CHATBOT + ASSISTANT
Some key features that we began to prototype through our AI platform were user verification, reporting content, connecting user profiles through shared interests, and a persistent prompt bar.
PROFILE INTERACTION
How the user interacts with the profiles of others was a crucial consideration for this product. Establishing a simple yet comprehensive pattern here meant the difference between meaningful and passive connections.
PROFILE: TESTING DIFFERENT PATTERNS
We went through many versions of thia screen to nail down the type of interaction we wanted users to have with each other.
BEING INTENTIONAL
Unlike other professional networking apps, we wanted the user's 'about' section to be large and pronounced at the top of the page. This way, you are prompted to learn about who someone is before you decide to interact with them.
BEGINNINGS
ToF offers starter prompts to get new and current users engaged. These can pertain to general topics or current events.
THE USER IS IN CONTROL
When ToF gets the information the user asks for, it puts the power back in their hands by allowing them to read for themselves, report the content as false or misleading, or save the content itself.
THREADS
Chats are saved and can be pinned for future use.
FINDING PEOPLE
ToF can provide articles, posts, blogs, etc., but it can also cross-reference authors with active account holders to find a match and get you connected with the source.
GROWING YOUR CONNECTIONS
The user can ask ToF to suggest other profiles based on the similarity of the content that they or others engage with.
SEAMLESS AI INTEGRATION
The persistent prompt bar allows the user the request information about whatever they are actively looking at.
USER PROFILE
Once a user verifies themselves by providing the necessary credentials, they will be able to report and verify content on ToF.
PERSISTENT PROMPT BAR
The prompt bar is always present, allowing users to ask questions, request information, and reference anything on the screen at any given moment.
MEASURING SUCCESS
Metrics that could be used are the amount of content verified vs. reported by verified users, links visited through the app, and public journalist reviews.
FUTURE CONSIDERATIONS
Future implementations include using AI to generate forums based on content and user-founded, owned, and operated blogs.
POTENTIAL BUILD STRUCTURE
As it took 2 designers to create this project, 2-3 designers would be expected to take part. In addition, a single PM and 1-2 UX researchers working closely with real journalists would provide invaluable insights.
LEARNINGS
This was my first hackathon experience. I made new friends and learnt a lot about user research and tending to user needs. I stepped far out of my comfort zone for this project, and I'm ever grateful for it.