“We’re drowning in a sea of fake news..!”
I wasn’t fully aware of the extent of the fake news problem until I participated in the Google-hosted APAC Trust Media Summit design sprint workshop in 2018. I was shocked to learn that journalists and newsrooms struggle daily to identify and discredit false information.
In Asia, the situation is especially dire, with a large number of social media users and a lack of mobile tools to detect fake news. This makes it even more challenging to separate fact from fiction.
I was part of a team of journalists, software engineers, students from NUS(National University of Singapore), and Google experts that designed and validated a mobile verification tool during the Google-led design sprint at the summit.
As a kick-off, journalists shared insightful presentations that gave us a glimpse into the challenges they face and their investigative processes.
Several key problem areas have been identified based on their sharing:
- the difficulty of translating text on images
- the need for speedy image verification. Sometimes, they need to select a specific part of an image or tweak it for investigation purposes, and this is not easy to do.
- the difficulty in accessing older sources for understanding the root of the image and context.
- the difficulty in accessing older sources for context and investigation purposes
- the lack of adequate tools for these tasks on mobile phones
We kicked off the design sprint by posing “How Might We” questions to generate broad solution concepts. Then, we moved onto the “Crazy 8's” exercise to flesh out detailed ideas. The outcome was a collection of innovative and fresh ideas, which we then narrowed down to the most impactful and value-adding ones through voting by the team.
With the ideas selected, we began crafting a mobile-centered web app.
We designed an interface that would allow users to quickly upload an image and perform a range of image-related tasks. It included features such as image flipping, cropping, rotation, and a set of filters to sort results based on date.
To address the issue of language translation, we integrated Google Translate API into the app, allowing users to translate text within images and choose from a range of languages.
With the positive feedback from journalists, we continued to refine the design, making it more efficient. Then tech team built a prototype in a few hours and worked on improving the performance.
With the aim of verifying our design hypotheses and discovering areas of improvement, I planned a quick user testing.
The primary validation points were to ensure that:
- users can upload an image in the app
- users select and rotate specific parts of the image for investigation purposes
- Users can translate text within the image
- Users can see limit results to the past 24 hours
- Users can effectively determine the authenticity of the image
The test demonstrated a high level of success in completing tasks. Additionally, we received valuable feedback from the participants. Journalists were especially impressed with the intuitive design of the interface.
But like as always, it didn’t tell us everything. We gained a wealth of valuable feedback from the participants.
- Identify a method for rapidly summarizing the crucial information that a fact-checker needs to know about their image.
- A few users found the cropping UI to be a bit finicky. Explore alternative options.
- Discover a solution to expedite processing time as it currently takes too long.
- Some users were unsure about the main purpose of the app. Consider incorporating an introductory flow to increase understanding.
- Some test participants suggested that having information about the location of each source would be useful.
- Participants suggested that having information on the degree to which an image has been manipulated would assist in their investigation, such as the percentage match to the original source.
- The possibility of verifying videos should also be taken into consideration.
Irene, the lead at Google News Lab, presented the results of our design sprint to a group of journalists, who were impressed by the tool’s ability to instantly analyze an image’s history, identify objects, languages, and provide automatic translations. It was a thrilling moment to hear their enthusiastic reactions.
After the 2018 summit, the Storyful team developed a web verification app named ‘Source’ and I was invited to assist with its usability testing. The tool maintained a similar structure and design as the prototype created in 2018 but with improvements such as a new skin and enhanced features.
I was grateful for the opportunity to participate in a compact and highly productive one-day design sprint organized by Google. It also allowed me to reflect on the potential of technology in addressing challenges such as fact-checking and promoting trusted media.
