We’ve all seen how bots and misinformation can spread false ideas like wildfire across the internet and social media platforms. Deciphering the genuine human voices from the cacophony of bots, trolls, and inauthentic accounts can be a daunting challenge.
When it comes to things like public health emergencies and national election cycles, this challenge can become a real problem with serious consequences. Similarly, it can make it hard for researchers and companies to collect accurate study data and authentically connect with their target audiences.
To combat this issue and cut through the noise, Northeastern researcher Professor Jason Radford and his team at the Social Design Lab are leading the True Voice Project — a pioneering research initiative to verify authentic social media accounts and revolutionize how we analyze online conversations.
This work has earned them selection as one of the Fall 2023 Spark Fund awardees.
Finding the Real Conversations
In 2015, Radford needed to find a way to ensure his social media research was based on real people’s behavior, not bots or inauthentic accounts. Unlike conventional approaches that focus on identifying bots and fake accounts, he decided to verify the legitimacy of actual users in data samples and discern genuine human behavior amidst the digital noise.
“When we’re looking at Twitter X or Facebook data or any kind of big social network, we don’t want to treat autonomous or inauthentic accounts as human beings,” says Radford. “We want to study what real people are doing and talking about.”
Radford and his team started by validating a sample of 20,000 accounts against publicly available data, such as the phone book, to identify accounts that could most likely be associated with real people. Over the years, they have slowly expanded their dataset to include 1.7 million real people verified on Twitter X (not to be confused with the “blue checkmark”) with a success rate of more than 99.998%.
This groundbreaking methodology enhances the scientific rigor of social media research and unlocks a wealth of possibilities for understanding human interaction in diverse contexts. From studying the dissemination of misinformation to analyzing trends in public opinion, there are broad implications for this research.
Of course, privacy concerns have been a central focus throughout the research process, prompting the team to adopt a nuanced approach to data aggregation and dissemination. By safeguarding the anonymity of participants while providing aggregated insights, they strike a delicate balance between transparency and privacy, guided by the principles of responsible innovation.
Looking ahead, the team envisions a future where validated accounts serve as the cornerstone of social media analytics, empowering individuals and organizations to navigate the digital landscape with confidence and integrity, and to pave the way for a more transparent, authentic, and meaningful online discourse.
Commercialization with the CRI
The True Voice project has potential applications spanning disciplines from journalism and politics to marketing and public affairs. Validated accounts pave the way for more accurate research and authentic engagement with digital communities by offering a more accurate and transparent representation of online discourse.
Therefore, the True Voice project seeks to transform their research findings into practical solutions for real-world challenges. By partnering with companies, governments, and organizations engaged in social listening and analytics, the team aims to provide access to high-quality data and empower decision-makers with actionable insights.
“We’ve worked with the CRI on several occasions to get help navigating tech transfer for a couple of projects, including True Voice,” says Radford. “Each experience with the CRI has helped us better understand the technology and what kinds of protections are available to us. Using their industry partnership connections, we also worked with the CRI to evaluate our readiness for an SBIR.”
Awards & Funding
Northeastern CRI’s Spark Fund partly funds this technology, as the team was selected as one of the Fall 2023 Spark Fund awardees.
“The Spark Fund offers a fairly simple way to get feedback on a translational idea,” says Radford. “It took a couple of submissions before we won an award, but each submission was a substantial improvement on the last. It definitely helped sharpen our ideas.”
The team will use the funds from the Spark Fund Award given by Northeastern’s CRI to create a demo to collect user feedback. They’ll use this demo to illuminate the strengths and limitations of the data and to better understand what practitioners are looking for.
This work builds on a series of research grants from the NSF.
Learn More
Learn more about Radford’s research and the five other Spark Award grantees here.
Written by Elizabeth Creason