The following teaching lesson was awarded first place in the 2016 AEJMC Best Practices in Ethics in an Emerging Media Environment Teaching Competition.
Given recent growth in social media technologies, it is increasingly difficult for journalists, educators, and students to keep up with ethical issues that arise from product development. It is important to be critical of how technologies come to market, and to be aware of technological bias and its impact on journalism and mass media communications.
This class activity explored bias in the development of digital technologies used to represent skin color. The activity draws on use cases from the early days of Kodak film to issues surrounding Twitter’s new “racially diverse emoji,” and how students can collaborate to make change.
The class activity was designed to expose students to the presence of bias in product development, specifically issues involving skin color and technology, and how bias impacts communications across various industries (e.g., journalists, marketers, advertisers, etc.). The activity draws on use cases from the early days of Kodak and biased color film, to issues surrounding Twitter’s new racially diverse emoji (Chowdhry, 2015) (“diverse emoji”), and how students can collaborate to raise awareness of ethical issues in communications and make change.
The activity began with a discussion of how some products that come to market are biased and not inclusive of people of color, starting with Kodak color film. To process color film, the company created a Shirley card, which was a photo of a white woman, to assist color lab technicians with developing color film. Kodak did not have cards for people of other races, which made it difficult to correctly print photos of people of color (Ali, 2015). We then discussed contemporary issues around bias in facial recognition, such as web cameras that lack sufficient technology to properly work for people of color, to illustrate how biased product development is not just a thing of the past (Albanesius, 2009).
As the number of white-skin emoji increased (Newton, 2014), celebrities and influencers began to raise ethical issues around the lack of diversity in emoji (Perez, 2014). In response, Apple released diverse emoji in their iOS 8.3 update on April 8, 2015 (Chowdhry, 2015). These emoji are created by applying an emoji modifier based on the Fitzpatrick Scale—a well-known scale for classifying human skin color based on how it reacts to ultraviolet light—to a default emoji (Warren, 2015). With this new release, the Unicode Foundation, which governs the release of emoji, has made efforts to standardize the default skin color as yellow, which caused some issues in the Asian community (Warren, 2015). On December 3, 2015, 232 days later, Twitter still had not updated their desktop computer application (“desktop”) to display them correctly. I presented students with the observation that Twitter’s desktop application was not able to properly display the new skin tone emoji that Apple released (see Appendix, Figure 1).
The rollout of diverse emoji was not coordinated amongst major companies such as Apple and Twitter, which resulted in people of color being marginalized. When non-white people created Tweets with diverse emoji on their mobile phones, they were frequently represented by a white emoji (not only the new standard default yellow) plus the skin tone swatch that they chose, when viewed on desktop (see Appendix, Figure 1). For example, if an African American person selected a new diverse emoji on mobile, it would display a white emoji plus a new Unicode Fitzpatrick swatch of their selected skin tone on desktop prior to the new iOS release.
People of color were further marginalized beyond the Twitter desktop application as journalists frequently embed live Tweets in major publications. Due to the interconnectedness of Twitter and digital publishers, any publication that embedded diverse emoji would have displayed them incorrectly as a base white/yellow emoji plus a new Unicode Fitzpatrick swatch to their desktop audience.
The lecture portion of the class reviewed how product development roadmaps and timelines may differ due to what companies prioritize, as well as how social media companies and journalists are interconnected.
In this class segment, I also covered how social media are used for social change and highlighted a new product called Thunderclap.it, which amplifies messages by allowing large groups of people to post messages on social media (e.g., Twitter, Facebook, etc.) at the same time. At the conclusion of the lesson, the class was invited to participate in an optional Thunderclap campaign where students had the opportunity to ask Twitter to prioritize updating their desktop application to display diverse emoji.
As journalists and product developers in training, students should be aware of how social inequalities are reproduced in products that we use, how this marginalizes people, and how ethical issues in one industry can impact others and reinforce issues such as institutional racism and oppression.
Students developed critical thinking around biased product development and Twitter’s product development priorities. For example, one student raised the issue that during the time that Twitter did not address the skin tone swatch issue, they prioritized changing the Favorite button from a star to a heart, an arguably trivial change for users.
The Thunderclap campaign (see Appendix, Figure 2) called on Twitter to prioritize emoji equality and garnered more than 45 supporters, including many students from the class, and achieved the potential to reach 56,928 users on social media with our message. Before the campaign ended, Twitter updated their product in line with the goals of the campaign.
Albanesius, C. (2009). HP Responds to Claim of ‘Racist’ Webcams. Retrieve from http://www.pcmag.com/article2/0,2817,2357429,00.asp
Ali, M. (2015). A brief history of color photography reveals an obvious but unsettling reality about human bias. Retrieved from http://www.upworthy.com/a-brief-history-of-color-photography-reveals-an-obvious-but-unsettling-reality-about-human-bias
Brownlee, J. (2015). Apple Starts Adding Racially Diverse Emoji To The Mac. Retrieved from http://www.fastcodesign.com/3042175/fast-feed/apple-starts-adding-racially-diverse-emoji-to-the-mac
Chowdhry, A. (2015). Apple Releases iOS 8.3 To The Public, It Has New Emoji. Retrieved from www.forbes.com/sites/amitchowdhry/2015/04/08/apple-releases-ios-8-3-to-the-public-its-the-update-with-the-new-emojis/
Newton, C. (2014). Give your friends the finger when these new emoji arrive on your phone. Retrieved from http://www.theverge.com/2014/6/16/5815558/give-your-friends-the-finger-when-these-new-emoji-arrive-on-your-phone
Perez, S. (2014). Proposed Changes To Emoji Standard Would Allow For More Diversity, Increased Selection Of Skin Tones. Retrieved from http://techcrunch.com/2014/11/04/proposed-changes-to-emoji-standard-would-allow-for-more-diversity-increased-selection-of-skin-tones/
Warren, C. (2015). How the new Apple emoji got their skin tones: It’s not what you think. Retrieved from http://mashable.com/2015/02/26/diverse-emoji-explainer/
Figure 1: Twitter Skin Tone Emoji Issue on Desktop
Here is an example of a Black Lives Matter activist using a black fist emoji. Prior to the update, the black fist was displayed as a white fist plus a new Unicode Fitzpatrick swatch, as the desktop version was not configured to display diverse emoji.
Tweets are frequently embedded in journalists’ digital work and the Fitzpatrick Swatch emoji was seen in top tier digital publications that were using Tweets as sources.
Figure 2: Thunderclap Campaign Page