Memes are used by many people to spread mis- and disinformation, or as a way to engage in politics and current affairs, so it is important for journalists to not just cover their virality or truthfulness, but to also understand the communities who create and share these visuals.
A panel at the International Journalism Festival in Perugia today (12 April) discussed the challenges of identifying the various types of visuals that are intentionally or unintentionally used to spread false information, as well as how journalists can present the broader context behind the origin of memes.
Farida Vis, director of Visual Social Media Lab, presented some preliminary findings from an upcoming research produced with First Draft into how visuals were used to spread misinformation on social media during the UK and French elections in 2017.
The project, which originated from a session held at the festival in 2017 to look at the origins of memes and their role in shaping conversations both online and offline, looked at a total of 95 visuals covered by verification collaboration CrossCheck during the elections in the two countries. The research identified 21 types of visual formats used, and analysed their level of accuracy, topic, motivation and source, as well as who had posted the visual and where.
Building on the seven types of mis- and disinformation content First Draft had previously identified, including fabricated content, false context and misleading content, the team found three more categories that can be applied specifically to images:
- no context provided: where the origin of a visual cannot be identified so it is harder to prove that it's inaccurate;
- exaggeration: particularly with statistics, where the information accompanying an image is "based on a real trend but is exaggerated or played down in order to deceive";
- genuine mistake, where the image contains misleading information caused by a genuine error.
Out of the 95 visuals covered, the most frequently shared formats were videos (20), followed by photographs (17), data visualisations (9), photographs with text (7) and screenshots of tweets (5), and people were mostly spreading material that fell under the 'false context' category, Vis said, where 'genuine content is shared with false contextual information'.
"[Fale context] is what's incredibly hard to deal with algorithmically when we talk about verification, because you are not dealing with images that are technically traceable as false. The image isn't the problem, it's the context, and you can potentially have different occurrences every single time."
An Xiao Mina, who leads the product team at Meedan and is working on a book about memes and global social movements, said that as journalists are working to identify problematic content, "in addition to thinking about the information, we should also be thinking about issues around the emotion, identity and narrative" that memes tap into.
"Maybe contemporary fandom is a more useful framework for some contemporary politicians' media strategies.
"In other words, we need to start taking Kim Kardashian, The Walking Dead, more seriously to understand why these memes are popular and how people are using these memes to engage in politics," she added, pointing to the similarities between brands that transcend their original format and become part of people's online and offline lives, and public and political figures such as Oprah or Trump.
Presidents and people in power "have to adapt to people's communication styles", she added, something Trump has been embracing by sharing memes, joining into people's jokes and conversations around his 'covfefe' tweet, and getting his audience to talk about his upcoming moves or meetings by teasing trailer-like information on social media.
"The ability to embrace people's tendencies to want to talk about you is a powerful way to draw attention. When Trump retweets memes, we should think about the role journalists play in amplifying these messages, and about what happens when they become part of first page news," she said, explaining that the stories of the communities who produce and share memes should be told as part of the broader context to avoid "unintentionally amplifying misinformation".
Jennifer 8. Lee, co-founder and chief executive of Plympton, who moderated the discussion, said that journalists tend to see a meme spread and focus on covering its virality, but "the meme is the manifestation of something bigger" and journalists should also observe the wider implications of the cultural phenomenon behind the visual.
"You tend to get things that push the button emotionally – religious tension, homophobia, that's what [the context] will be focused on," Vis said. "Anything that is a metaphor or that alludes to threats by playing on the imagination of fear works much better than cold hard facts."
"Many of us live with a narrative we grew up with about society and what's fair," Xiao Mina added, "so when you're looking at these images in addition to thinking about their truthfulness, we need to understand where people's emotions, which are based on cognition and beliefs, come from."
Memes can also be a potential source for stories, Vis explained, flagging to reporters that there are certain audiences whose voices aren't included "and not necessarily because they have nothing to say", so journalists should find ways to collaborate with each other and build partnerships in order to tap into these groups and tell their stories.
Free daily newsletter
- The impact of AI on journalism and democracy
- 2019 Reuters Digital News Report finds that trust in the media continues to fall
- Computational photography can enhance pictures but also produce fake news images
- Tip: Use these quick verification shortcuts
- Full Fact secures $2 million Google AI grant to fight misinformation