Misinformation circulating online can be an issue to tackle. On social media many such pieces of misinformation and conspiracy theories are circulated every now and then. These create an unnecessary amount of panic and fear amongst the audience. Consuming such posts carrying misinformation may lead to even more health hazards.
Using Social Media to Track Conspiracy Theories
Video Credits: Los Alamos National Lab, YouTube
A postdoctoral researcher in the Information System and Modeling Group at Los Alamos National Laboratory, Courtney Shelley, who is also the co author of the study which was published in the ‘Journal of Medical Internet Research’ said, “A lot of machine-learning studies related to misinformation on social media focus on identifying different kinds of conspiracy theories. Instead, we wanted to create a more cohesive understanding of how misinformation changes as it spreads.
She fruther added that, “Because people tend to believe the first message they encounter, public health officials could someday monitor which conspiracy theories are gaining traction on social media and craft factual public information campaigns to preempt widespread acceptance of falsehoods.”
According to the study the tweets which contain false information have more negative sentiments incorporated in them compared to those tweets carrying real facts and information. This then gives rise to conspiracy theories which make use of such details.
Image Credits: Los Alamos National Laboratory
In a study which was titled ‘Thought I’d Share first’ made use of the publicly available data on Twitter in order to characterize four conspiracy theory themes related to Covid 19. With this study they concluded and could read out trends such as 5G cell towers spread the virus, the virus being bio engineered etc.
Dax Gerts, the study’s co author and a computer scientist in Los Alamos’ Information System and Modeling Group said, “We began with a dataset of approximately 1.8 million tweets that contained COVID-19 keywords or were from health-related Twitter accounts.”
They further added that, “From this body of data, we identified subsets that matched the four conspiracy theories using pattern filtering, and hand labeled several hundred tweets in each conspiracy theory category to construct training sets.”
With the help of the data collected from each of the four conspiracy theories the team was able to build random forest machine learning or artificial intelligence (AI) models which were capable of categorising the tweets regarding Covid 19 as misinformation or not. This also allowed the researcher’s team to observe the way individuals speak about such conspiracy theories online and also on social media. And also facilitated to observe the changes happening over the period of time.
Upon further research the study found out that with the help of a supervised learning technique conspiracy theories could be automatically identified. Futhermore with the help of an unsupervised learning approach (dynamic topic modeling) it can be possible to explore changes in word importance among topics within each theory.
Courtney Shelley said “It’s important for public health officials to know how conspiracy theories are evolving and gaining traction over time. If not, they run the risk of inadvertently publicizing conspiracy theories that might otherwise ‘die on the vine.’ So, knowing how conspiracy theories are changing and perhaps incorporating other theories or real-world events is important when strategizing how to counter them with factual public information campaigns.”