Studies have shown that news with scientific-sounding content is trusted more than other types. Therefore, any misinformation in the scientific domain can cause significant public risk as was evidenced during the recent COVID-19 pandemic. Our lab has a long history of working on mis/disinformation. In our recent paper (presented today at the #AAAI Workshop on Preventing and Detecting LLM Misinformation), we explore the problem of automatically detecting scientific misinformation in the wild, while also presenting relevant evidence material from scientific publications to backup the findings. Since there is no other labeled datasets available, we also provide a balanced datasets for anyone interested in trying it out. Our paper can be found at: https://lnkd.in/e7qcaZqZ Our dataset can be downloaded from: https://lnkd.in/e2mYpmFt Our earlier work and dataset in a related area (ASONAM '21) can be found here: https://lnkd.in/e3Yxnpgv Congratulations to the student team who made this happen! Yupeng Cao, Aishwarya N., Elyon Eymife and Nastaran JP Soofi! #AAAI PDLM
This is an important and timely contribution to addressing scientific misinformation, especially in an era where trust in information is critical. Congratulations to the team on this great work!
Worthwhile effort- especially in these times!