Beyond (Mis)Representation: Visuals in COVID-19 Misinformation.

Area of Science:

• Medical Visual Communication
• Digital Health Misinformation Analysis
• Information Science

Background:

• Visuals are increasingly used in health misinformation, particularly concerning COVID-19.
• Understanding the role and function of these visuals is crucial for combating false narratives.
• Existing research often overlooks the specific ways visuals are employed in misinformation campaigns.

Purpose of the Study:

• To analyze the frames and functions of visuals used in COVID-19 misinformation.
• To identify how visuals are employed to illustrate, evidence, or impersonate authority in false claims.
• To differentiate between manipulated and mislabeled visuals in misinformation.

Main Methods:

• A mixed-methods analysis was conducted on 96 examples of COVID-19 visuals rated false or misleading.
• Visuals were analyzed for their framing and functional roles within misinformation.
• Examples were categorized based on manipulation techniques, distinguishing simple alterations from advanced AI methods.

Main Results:

• Six distinct frames and three primary functions of visuals in misinformation were identified.
• Over half of the analyzed visuals served as explicit evidence for false claims, primarily through mislabeling.
• A small number of manipulated visuals were found, all created with basic tools, with no instances of AI-generated deepfakes.

Conclusions:

• Visuals play a multifaceted role in COVID-19 misinformation, extending beyond simple representation.
• The prevalence of mislabeled visuals underscores the importance of source verification and contextual analysis.
• Future research should consider a broader scope of visual functions and analysis techniques beyond representational aspects.