Written evidence submitted by Dr Jingrong Tong (SMH0025)

PhD, Senior Lecturer, School of Journalism, Media and Communication, University of Sheffield

UK Parliament, Science, Innovation and Technology Committee, “Social media, misinformation and harmful algorithms”

Introduction

1. I study social media communication, and social media content moderation practices and policies and would like to share a few perspectives on which bodies should take the responsibility for spreading untrue and harmful content. I have published widely on this topic, including the following recent publications:

• Tong, J. and Zuo, L. (2018) Tweeting the Environment #Brexit. London: Emerald Publishing Limited.
• Tong, J. and Zuo, L. (2020) Mainstreaming populism through the Twitter practices of politicians and the news media: a case study of the 2016 Brexit referendum debates, in Information Polity., 25, 361–375.
• Tong, J. and Zuo, L. (2021) The Brexit referendum on Twitter: a mixed-method, computational analysis of how Twitter works in shaping political discourse, London: Emerald Publishing Limited.
• Tong, J. (2022) “Artificial Intelligence (AI) is not a panacea: policing content on social media platforms, three dilemmas and their ethical and legal implications”, in AI and the Media, edited by Pihlajarinne, T. and Alen-Savikko, A., Edward Elgar Publishing.

2. Drawn from these recent publications, my response is focused on addressing the following aspect: “Which bodies should be held accountable for the spread of misinformation, disinformation and harmful content as a result of social media and search engines’ use of algorithms and AI?”

3. My research demonstrates:

• The responsibility of and dilemmas faced by tech giants in cleaning and delivering healthy content amid technological, linguistic limitations and tensions with commercial interests.
• The importance of government regulations in balancing content moderation, delivery, and freedom of expression
• The accountability of social media users as platforms amplify social emotions, especially those tied to polarised ideologies, fuelling the spread of false and harmful information
• The responsibility of elites and influencers who often dominate asymmetrical information flow on social media platforms and promoting specific ideologies

4. The responsibilities and dilemmas faced by tech giants. Social media companies use algorithms and AI for various purposes, primarily content moderation and personalised delivery to users. In the UK, as in other democracies, the government is not directly involved in controlling online speech. Social media platforms are responsible for maintaining a healthy online environment, setting their own rules, and addressing harmful content. However, my research indicates significant challenges in effectively moderating harmful content. Technological and linguistic limitations hinder automated systems, necessitating human. Even with advanced AI tools, human moderators remain essential, though their effectiveness is limited by the vast volume of content on social media platforms. Moreover, accurately detecting harmful content within images and videos remains particularly challenging. In addition, as businesses, social media companies prioritise their commercial interests. This can result in discrepancies between their content moderation practices and the best interests of online speech and democratic discourse. For example, their use of algorithms and AI to rank content by popularity and tailor it to user interests reflects their commercial priorities and business models, often conflicting with their responsibility to uphold healthy online speech. The prioritisation of popularity and attention, often central to these algorithms, can contribute to the spread of misleading or harmful information.

5. The importance of appropriate government regulations. Given the increasing issues of fake news and hate speech, allowing social media content to go unchecked is irresponsible. However, social media companies are private businesses after all and cannot be fully relied upon to moderate and deliver content.  Regulations and policies are essential to hold social media companies accountable. However, government regulations must recognise the practical challenges social media companies face in addressing false and harmful information, while also balancing content regulation with the preservation of freedom of expression. Excessive regulation could lead to censorship. 

6. The accountability of social media users. Social media platforms have become powerful amplifiers of social emotions, particularly those associated with polarised ideologies. When confronted with contentious issues, social media users often find it easy to align themselves with the side they agree with, leading to ideological antagonism and creating an environment conducive to the spread of false and harmful content. For example, my research reveals the prominence of ideological antagonism and confrontation in the period leading up to the 2016 Brexit referendum.

7. The responsibility of elites and influencers. My research reveals that within the vast ocean of information on social media platforms, most messages fail to capture users’ attention. A small fraction of content, typically those garnering retweets or likes, dominates the discourse surrounding specific topics. The majority of messages remain largely unnoticed and unengaged with by other users. Within this landscape, a distinct asymmetry emerges: elite users, such as politicians, CEOs, journalists, prominent activists, and influential community members, command a disproportionate share of user attention. Information often originates from these elite accounts and rapidly disseminates through user forwarding. These influencers effectively set the agenda and dominate the flow of information within social media. While ordinary users contribute the majority of content, their messages rarely achieve significant circulation. This dynamic fosters an asymmetrical power structure, where elite users, often characterised by high popularity despite limited activity, exert significant influence over the platform. Conversely, ordinary users, despite their active participation, struggle to gain traction. The social media practices of elites and influencers can have profound consequences, including the amplification of specific political agendas and the dissemination of misinformation targeting marginalised groups such as immigrants and ethnic minorities. For instance, my research demonstrates the role of social media platforms, particularly X (formerly Twitter), in mainstreaming populist discourse through the social media practices of politicians and news media during the period prior to the 2016 Brexit referendum.

Recommendations

8. Strict regulations are necessary to govern the use of algorithms by social media companies and ensure the presence of healthy content on their platforms. However, regulatory expectations for content moderation must be realistic, recognising the inherent practical challenges. At the same time, these regulations must be carefully crafted to avoid undue restrictions on freedom of expression.
9. Meanwhile, educational campaigns for social media users are essential to help them understand what they can or cannot post, believe, or share, along with the potential consequences.
10. Identifying and engaging with elite members and influencers about responsible social media practices is essential.
11. I recommend a collaborative accountability model involving the state, social media companies, elite members, influencers, and ordinary users working together.
12. I also recommend involving independent regulators to oversee social media companies' content moderation, ensuring minimal state interference to prevent censorship while keeping in check the growing power of these companies.

17 December 2024