Skip to main content
SearchLoginLogin or Signup

The Promotion of Gender Equality on Social Media: Freedom of Expression and Challenges Posed by Artificial Intelligence

Tech companies that own social media services employ and deploy AI content curation and moderation for their commercial purposes. The operations of the AI tools often have negative effects on freedom of expression to promote gender equality.

Published onFeb 02, 2023
The Promotion of Gender Equality on Social Media: Freedom of Expression and Challenges Posed by Artificial Intelligence

The Promotion of Gender Equality on Social Media:

Freedom of Expression and Challenges Posed by Artificial Intelligence1

Jompon Pitaksantayothin


Social inequality is a common phenomenon which can be understood from sociological perspectives at the level of general theories (Bottero, 2005, 15-67). However, given the ideal world of equality for which the United Nations aims (OHCRH, n.d.), it is arduous to accept the harsh truth that billions of people are still desperately struggling against various forms of discrimination and biases. Gender inequality is a form of social inequality. As far as gender inequality in relation to women is concerned, “women are disadvantaged relative to similarly situated men” (Lorber, J., 2010, 4). A number of women suffer from gender inequality against them in many different societies around the globe. Nonetheless, the advents of the Internet and, more recently, social networking and social media platforms have provided unprecedented opportunities for the underprivileged groups of women to the exercise of their right to freedom of expression on the Internet. On these online platforms, women can tell and share their stories, express their ideas and opinions, and even campaign or organize social movements to call for the improvement of their lives and social status through democratic mechanisms. While they have more freedom of expression than they did in the pre-Internet era, the employment and the deployment of Artificial Intelligence (hereinafter “AI”) to curate and moderate content technology companies (hereinafter “tech companies”) have become a great hindrance to the full utilization of their freedom of expression to promote gender equality on the social networking and social media services.

This chapter examines this particularly interesting issue. It is divided into three sections. The first section concentrates on the building a conceptual framework which is based on the relationship between freedom of expression and democracy and the way in which the Internet and social media enhance the exercise of freedom of expression. Then the conceptual framework is applied to the promotion of gender equality on social media. The second section examines the challenges to freedom of expression in relation to the promotion of gender equality on social media which are imposed by AI content curation and moderation. In the final section, the two-tiered approach, i.e., policies and law, is proposed to deal with the problem.

Freedom of Expression, Democracy, the Internet (Social Media) and Gender Equality in Relation to Women

In this section, a conceptual framework is built based on freedom of expression, democracy, the Internet and social media. This section serves as a starting point for the following discussions in the remainders of this chapter. In the first subsection, the relationship between the right to freedom of expression and democracy is examined from a theoretical perspective to provide a general idea of the important roles which the right to freedom of expression plays in democracy. Then, in the second subsection, the focus is on how the Internet has provided the new spaces where Internet users around the world can exercise their right to freedom of expression, facilitating public discussions on social and political matters. After that, in the third subsection, the discussion is about how social media and social networking platforms are highly influential, attracting more people to engage in public discussions and campaigns which, in certain cases, have gain momentum and become social movements eventually. Lastly, the conceptual framework is applied to the issue of gender equality particularly in relation to women.

Freedom of Expression and Democracy

Undisputedly, the right to freedom of expression is considered as one of the most significant fundamental human rights. It is the right which gives people a legal claim against the state restrictions on the exercise of their rights to hold, express and seek ideas, opinions and information freely (Zeno-Zencovich, 2008, 1). At the international level, it is recognized by strongly influential international human rights instruments, such as the Universal Declaration of Human Rights (UDHR) and the International Covenant on Civil and Political Rights (ICCPR).

Article 19 of the UDHR provides:

Everyone has the right to freedom of opinion and expression; this right includes freedom to hold opinions without interference and to seek, receive and impart information and ideas through any media and regardless of frontiers.

Article 19 (2) of the ICCPR states:

Everyone shall have the right to freedom of expression; this right shall include freedom to seek, receive and impart information and ideas of all kinds, regardless of frontiers, either orally, in writing or in print, in the form of art, or through any other media of his choice.

At the national level, it is also guaranteed by national constitutions of many countries across the globe. One of the most well-known free speech (freedom of expression) clauses is the First Amendment to the United States Constitution.

The First Amendment reads:

Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof; or abridging the freedom of speech, or of the press; or the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.

The right to freedom of expression is a pivotal element of a healthy democracy. Before going any further, it is necessary to understand the basic of concept of democracy. The famous statement “Government of the people, by the people, for the people,” (Pasquino, 2008, 15) made by the United States President Abraham Lincoln in 1863 could encapsulate the whole idea of democracy. However, to make it easier to understand, it could be said that, in principle, democracy is a political system which has the people at large as its sovereign (Lively, 1975, 8). Moreover, the people have the power to control their government’s operation with regard to political matters and public or social policies through public decisions which are made either directly or by their representatives on the basis of equal political rights (Weale, 2007, 18).

The explanation of how freedom of expression plays a role in the functioning of democracy was provided by Alexander Meiklejohn, an English philosopher and an advocate of freedom of expression (Cram, 2002, 11). As mentioned above, the operation of a democratic government relies significantly on public decisions which derive from intelligent voting (Cram, 2002, 11). The intelligent voting could derive from “[public discussions] on an equal and inclusive basis, which deepens participants’ knowledge of issues, awareness of the interests of others, and the confidence to play an active part in public affairs” (Saward, 2000, 5).

Therefore, as a matter of principle, the people (as voters) should be able to gain access to a full range of ideas, opinions and information concerning public policies on political and social issues and interests of other members in their society. Freedom of expression allows the full range of ideas necessary for the voting to be available to the voters (Schauer, 1982, 38). Then, by considering upon and discussing about all ideas and all possible choices, they would be able to make intelligent voting on particular policies on political and social matters (Schauer, 1982, 38). Put differently, it could be argued that, without freedom of expression, the people would not be able to acknowledge and learn about various ideas of others for their consideration. The insufficient availability of necessary ideas, opinions and information would obstruct them from making intelligent political and social choices. As a result, it would be very difficult to expect that such political and social choices could be translated into good public policies.

The Internet as the New Spaces for the Exercise of the Right Freedom of Expression

Originally, what is known as the Internet today was developed for a military purpose during the Cold War era due to the US government’s ambition to create resilient communication networks that could continue functioning even in the case of a nuclear attack (Ryan, 2010, 11-22). As a result, the Advanced Research Projects Agency Network or ARPANET was developed (Yar and Steinmetz, 2019, 7). In the 1970s, the foundational network of computer networks initially laid down by the ARPANET was gradually expanded and used by other non-military organizations, such as some universities (Yar and Steinmetz, 2019, 8). Then, in the mid-1990s, the Internet, network of computer networks, was commercialized leading to the proliferation of commercial Internet Service Providers (ISPs) and the availability of the Internet to the general public (Yar and Steinmetz, 2019, 8). Several decades after its birth, the Internet has become an integral and indispensable part of people’s everyday activities, especially those in the younger generations (See generally, Mcmillan and Morrison, 2006). According to Statista (2022), as of April 2022, there were five billion Internet users around world, accounting for 63 percent of the world population.

While people have been using the Internet for various reasons, for instance, entertainment, online transactions, shopping and etc., it is obvious that the Internet has also been used as a medium of communication for the dissemination of ideas, opinions and information to others, as well as for participation in discussion groups (Barendt, 2005, 451). This is due to the fact that the Internet allows ordinary users (who could be anyone) to make their messages and contents available to “a much wider audience at a much greater speed …” (Rowland, 2005, 55) with much cheaper costs. The space on the Internet where people can express their views and thoughts, at present, comes in the forms of the traditional platforms, e.g., newsgroups, emails and web boards, and the more recent ones, namely, social networking websites and social media. As of April 2022, social networking and social media services had 4.65 billion users around the world (Statista, 2022).

On these online platforms, the sharp line which once rigidly separated information provider (the speakers) from the audience (the listeners) has become blurred, as the participants on both sides of the communication channel can shift “easily from one role to another” (Rowland, 2005, 56). User A posts his thought on a particular social issue on his social media wall. User B reads it, and then posts her opinion to respond User A’s post. Shortly afterward, User A posts his own opinion in response to User B’s view. In this example, originally User A was a speaker (a content provider) who initiated a conversation, while User B was a listener. However, User B became a speaker when she responded to User A’s post. Likewise, when User A read User B’s comment, he was a listener, but became a speaker again when he posted a message to discuss with User B. In this sense, User A and User B fluidly changed their roles between a speaker and a listener back and forth. More interestingly, both User A and User B can expand their interpersonal conversation to other users or even to the general public by sharing the post, allowing those people to take part in their original conversation. When more people have joined in, the post which was intended to be merely an interpersonal conversation at the beginning would gradually expand and become a discussion in which anyone can participate (excepting that it is set to be private). In this regard, it could be argued that the Internet facilitates public discussions to happen more frequently and easily. This appears to be in line with the view of Eric Barendt (2005, 451), a famous scholar on freedom of expression, that the Internet certainly gives people “much more equal opportunities for communication than the traditional press and broadcasting media”. In a democratic society, the more inclusive discussions on social and political matters in a free and open atmosphere are, the more likely for the public to make better choices on public policies.

Social Media and Social Movements

Facebook, YouTube, Instagram, Twitter and TikTok are some familiar names of social networking and social media platforms in the present time (Auxier and Anderson, 2021). Although they are different from each other in terms of design, function, purpose and target groups of users, what they have in common is that all of them provide space for conversation and, on certain platforms, public discussion. There are several unique features of social networking and social media platforms which attract a number of people to become their users and enthusiastically participate in the discussions appearing on their feeds. Firstly, they are a one-to-many medium of communications which makes it possible for a single user to disseminate his/her idea or information to many others, as well as a many-to-many medium of communications which allows many users to share opinions and information to many other users concurrently (Blossom, 2009, 31). Secondly, social networking and social media websites are a peer-to-peer channel of communications. The users who impart information are neither the controllers of the technology and the owners of social media platforms that they use, nor the state authorities who are in the positions with a power to command and control (Blossom, 2009, 31). There is not a hierarchy of status on the platforms among the readers or viewers. In other words, both the speakers and the audience are merely the users of the platforms. The atmosphere and the feeling that everyone is equally a peer in this way could make users feel more comfortable and have more confident to join in and express their views frankly and straightforwardly in the discussions. Lastly, perhaps most importantly, the views and the information imparted via social networking and social media platforms could greatly influence a number of people, especially the users. Due to the scalability, social networking and social media websites can be expanded to accommodate the users regardless of how many there are and will be; as a result, the content which was originally intended to share to a few users could have “a sudden influence a worldwide audience” (Blossom, 2009, 32). This is because the delivered content could reach millions of like-mined users around with just a click. The more the content is shared to the like-minded people, as time passes, the more influential it would be. As John Blossom (2009, 32) explains, it works in a similar way as campfire stories which were the stories of the past that have been repeatedly told from generation to generation, they could become powerful (as well as influential) legends.

Given what is argued above, it would not be an exaggeration to say that social networking and social media platforms serve as online fora where social and political matters can be discussed, sparking campaigns which could raise public awareness of particular issues or even lead to rallies demanding for social and political changes. As regards the role of social media in political movements, the most prominent example is Arab Spring in 2011, especially the protests in Tunisia and Egypt (Stepanova, 2011, 1). In the case of the Jasmine Revolution in Tunisia, Facebook allowed Tunisians to have a virtual public forum to discuss, express and share ideas, opinions and information concerning social problems of the country and the repressive regime of then the Tunisian President Zine al-Abidine Ben Ali; and this “[created] a common cause and understanding that kept mobilizing Tunisian ‘netizens’ to reclaim their rights as citizens, and, in the end to oust Ben Ali’s regime” (Muller and Hubner, 2014, 20). In the case of protests in Egypt, Twitter was used by free speech activists to disseminate information about the violence and brutality which the Egyptian government used against the protesters (Henderson, 2013, 5), leading to the overthrown of the governments of Hosni Mubarak. A clear example of the social media’s role in social movements can be seen in the #BlackLivesMatter movement 2014. It was ignited, inter alia, by the case of Michael Brown who was tragically shot to death by a white police officer in Missouri and the case of Eric Garner who was suffocated to death during his arrest by white police officers in New York respectively (Carney, 2016, 181). The aftermath of their death was a number of Twitter users attached #BlackLivesMatter to their tweets. Although the hashtag had been used for different purposes, the majority of Twitter users who attached it wanted to express their supports or “positive reference” (Anderson, 2016) to public discussions about racial problems and “a broader social movement” (Anderson, 2016).

It is true that not all campaigns and online movements end up with a sudden success, especially in the form of dramatical changes in political or social situations. However, it would not be wrong to say that social networking and social media services have a significant role in providing spaces for people to raise political and social issues and to discuss them, which could pave way to social and political movements both online and offline. Although social media users who participate in public discussions and campaigns may not achieve what they aim for at that time, the free and open space for discussions and activities on social media could plant a seed for further movements which could cause social and political changes sooner or later.

The Application of the Conceptual Framework to Gender Equality in Relation to Women

A conceptual framework that can be derived from the above discussion is that freedom of expression is an important human right that is a key element of a democratic society, as it allows all members to discuss political and social issues which could become good public policies as they could benefit all eventually. The Internet provides new space where users can exercise their rights to freedom of expression through playing roles as both the speakers to a wide audience and audience who could access to a wide range of different ideas and opinions from various speakers, encouraging more public discussions. Social networking and social media platforms facilitate people, especially those who have common interests in the same social and political issues, to engage in the inclusive public discussions, which, in turn, would become a strength to demand for changes.

As far as gender equality is concerned, freedom of expression and an ability to exercise it empowers women to deal with gender inequality through addressing the problems of discrimination that they are facing, voicing their views and ideas, organizing and participating in public discussions on how to liberate them from restrictions imposed by patriarchal ideology and to improve their social status to be equal to men (Article19, 2020, 6). Based on their right to freedom of expression, they could go further to call for other rights, for example, “the right to vote, the right to control their own bodies, the right to [unionize] and the right to equality before the law” (Article 19, 2020, 6). As they could unite, their voices would become louder, more persuasive and more influential. As regards public policies, it is obvious that gender inequality is a political and social problem that requires good and effective public policies. To attain good public policies on gender equality, it is necessary that all member in a society acknowledge and are aware of its importance, and ultimately make an intelligent choice by means of voting for the public policies that could promote, strengthen and sustain gender equality in the society. In this regard, it is significant that women need to know their rights to make an intelligent voting to prevent patriarchal restrictions. The right to freedom of expression plays an important role here by making it possible for them to access information and ideas of others from which they can learn to understand the rights that they deserve in general; which, in turn, would equip them with sufficient knowledge to “push for change, participate in decision-making” (Article 19, 2020, 6) through intelligent voting as argued above.

The Internet has been regarded as a new platform where people can express and disseminate their views to wide audiences around the globe. Furthermore, it allows the users to interact and exchange opinions with each other, creating online fora for public discussions. As contended above, the Internet facilitates public discussions and brings people together to partake in the discussions. It is also the case for the issue of gender equality. In the online discussion groups, regardless of where the speakers and the audience are in the world, women could express their ideas and thoughts, share information with and learn from other women. The more robust the public discussions to address the problem of gender inequality, the louder the voices that would reach the governments. Given this, it could be said that online public discussions could be turned into immense pressure which women could put on their governments to make and implement public policies which take into account gender equality.

As stated above, social media play a vital role in the organization of public discussions and campaigns on various political and social topics. Various platforms of social media have provided women with an unprecedent powerful channel to call attention of the world to focus on the issues regarding gender inequality against women in their countries, as well as opportunities to “[organize], protest, raise awareness of discrimination” (Article 19, 2020, 6). This could urge the policy makers to make a commitment to improve the situation of gender inequality in their countries (Loiseau and Nowacka, 1) in the end.

The #Metoo movement in 2017 can serve as a prominent example of how women have exercised the right to freedom of expression on social media platforms in relation to the problem of gender inequality against women. In October 2017, New York Times and New Yorker reported about a number of sexual abuse cases allegedly committed by Harvey Weinstein, an American film producer, prompting the media to investigate more deeply and reveal more cases of sexual misconducts by powerful men in different businesses and circles (Tuerkheimer, 2019, 1147-1148). At the same time, an American actress, Alyssa Milano, posted on her Twitter account: “If all the women who have been sexually harassed or assaulted wrote ‘Me too’ as a status, we might give people a sense of the magnitude of the problem on her twitter.” (Milano, 2017)

As an aftermath of her bold call, the attachment of #Metoo on social media posts went viral and catalyzed a huge online movement, one that demanded more proper treatments of women, to spread globally. Catharine MacKinnon, a reputable scholar and a feminism activist, interestingly called this movement “butterfly politics”, which referred to a phenomenon whose the starting point was a simple action (which in this case was the post of #Metoo by Alyssa Milano), but rendered an unexpected and enormous impact on societies and politics (which in this case was a huge online movement leading to “[the] shifting [of] law, cultures, and politics everywhere”) (MacKinnon, 2020, 1). A number of positive changes to more equal treatments to female workers ensued. Some examples that can be seen in the United States include the prohibition of the agreements between male employers and female employees that the employees would not disclose sexual harassment, the protection of more various groups of workers (e.g., independent contractors), the Congress’s reform of sexual harassment report process and financial restitution offers to some victim-survivors (North, 2019).

Unquestionably, the #Metoo movement provided a channel for women who were sexually harassed and abused to speak out and call other sexual victim-survivors to do the same in order to gain public attention to the problem of sexual violence against women particularly in male-dominated environments and cultures (Fileborn and Loney-Howe, 2019, 11). The voices, thoughts, opinions and information from the victims and the victim sympathies had gained a momentum to increase public discussions on gender inequality-related issues, namely inappropriate behaviors of men towards women and, more importantly, the imbalance of power between men and women (Khomami, 2017). This, in turn, led to rallies calling for changes for equality between men and women. It could be said that the #Metoo movement is perfectly in line with the application of the conceptual framework developed above to the issue of gender equality in relation to women.

Challenges Posed by AI to Freedom of Expression to Promote Gender Equality on Social Media

As discussed in the previous section, the Internet has empowered the women, who have experienced agony and unfair treatments caused by gender inequality, to voice their thoughts to millions of Internet users around the world, as well as catalyzing more public discussions on these gender inequality-related issues. Furthermore, social media have made their messages delivered directly to social media users, urging them to engage in the public discussions. As in the case of the #Metoo exemplified above, social media even made it possible for gender equality advocates to organize online campaigns and social movements demanding equal opportunities and social status for women. Given these, as a matter of principle, on the Internet women should have exercised their right to freedom of expression more effectively than ever before, and hence the situations of gender equality for women should have been improved to a great extent. Nonetheless, in reality, it is not always the case since there are still several challenges. AI technologies operating behind social networking and social media platforms could obstruct the opportunities for women to make the most of the right to freedom of expression on the Internet. Not only do these challenges make it more difficult for women’s efforts to progress, but also perpetuate the situation of gender inequality in societies. This section examines some of these AI-related challenges to understand how they negatively impact the promotion of gender equality for women through the exercise of the right to freedom of expression on the Internet.

I. AI Technologies and Social Media

Before going further, it is necessary to understand how AI technologies are normally employed and deployed on social networking and social media platforms. It is also crucial to note with an emphasis that, at the time of writing, social networking and social media services are owned and run by big tech companies for commercial purposes.

The advent of Web 2.0 has dramatically changed the way in which people use the Internet. While the technology of Web 1.0 allows Internet users only to read the pre-written content on websites, the Web 2.0 technology which has “read-write interface” (Stephens, 2009, 2) has made it possible for the users to create their own content and upload it to the Internet for sharing with other users with ease. The user-generated content (or user-created content) has substantially changed the ecosystem of the Internet by making it highly interactive. The online interactivity not only makes the Internet different from traditional media which have only a one-way communication capacity, but it also becomes the heart of the functionality of social networking and social media services. The highly interactive nature of social network and social media websites makes it easier for the users to employ user-generated content (such as textual posts, photos and video clips) to attract other users and befriending them online. It is not surprising that social networking and social media services are tremendously popular nowadays.

The great popularity of social networking and social media platforms around the world has leaded to exponential increases of user-generated content to be shared with and circulated among the other users. According to Bernard Marr (n.d.), a well-known writer in the fields of business and technology, on Facebook (which is the most popular social networking website at the moment), 293,000 statuses and 510,000 comments are posted every minute and over 300 million photos are posted daily. In addition, the infographic provided by Data Never Sleep 5.0 (n.d.) also give interesting statistics that, as of 2017, 456,000 tweets were posted on Twitter every minute; and 46,740 photos were posted on Instagram every minute. Given the sheer volume of data and information circulating on the Internet, it is beyond human’s capacities to analyze them before letting them appearing on the platforms (Llanso et al., 2020, 3). Therefore, many social network and social media companies rely on AI technologies to “manage information flows and to shape and arbitrate content online” (Bukovska, 2020, 19) Normally, AI technologies are used for these purposes by curating and moderating online content (Bukovska, 2020, 19).

According to Llanso et al. (2020, 14), AI content curation is the system which can organize a sheer amount of content items and “present (“curate”) a selection of content (“recommendations”)” to individual users as per their particular preferences or interests. The process of content curation begins with AI’s analyses of data collecting from how a particular user and other users who have preferences similar to him/her interact with particular content on the Internet, such as “website visits, articles read, social media behavior such as clicks and likes” (Llanso et al., 2020, 14). Then, based on these analyses, AI will match up the individual user’s preferences with the gigantic data bank of content on the Internet and then makes recommendations of specific content items to him/her (Llanso et al., 2020, 14). It can be said that the information that AI content curation selects and recommends is the personalized or custom-made information exclusively for that particular user. While it is undeniable that the AI content curation systems could prevent the users from being overwhelmed with a very large amount of content items, especially those which are irrelevant or beyond the scope of the users’ interests, it also the method from which the tech companies make profits, by “[generating] advertising revenue by increasing user engagement” (Llanso et al., 2020, 15).

The AI content moderation operates differently. Its main function is to analyze online content in textual, visual or audio forms, before making an automated decision whether the detected content is illegal or problematic, and thus is needed to be filtered out or taken down (Bukovska, 2020, 19). Regarding its mechanism, the AI content moderation will take down the content in question if it matches anything on its database of illegal or violative content. In the case of AI with machine learning technologies, the content will be removed in accordance with the predictions of AI (Helberger et al., 2020 A, 7). The results of prediction depend on the datasets of violative content which have been used to train AI (Helberger et al., 2020 A, 7). To a certain extent, the AI content moderation could prevent the users from harmful or legally prohibited content (Bukovska, 2020, 60), as well as safeguarding the technologies companies against any possible risks of civil and/or criminal litigations, since such the content suspicious of being harmful or unlawful would be filtered out (at the first place) or removed (later on), and thus do not appear on social networking or social media feeds of the users.

Given what just described, it could be said that the AI content curation and moderation are beneficial for both the users and the tech companies. However, it is very important to bear in mind that all of the aforesaid processes are automatically carried out by AI systems without human involvement. Therefore, what messages, information, ideas and opinions which are allowed to be shown on the users’ feeds are automatically decided solely by AI, whose operation depends solely on the designs of their algorithm. Furthermore, it is important to note that social networking and social media platforms are owned and run by the private sector, namely the tech companies. These companies have an absolute power to set their own AI algorithms to sever their commercial interests. Moreover, they also have an authoritative power to prioritize what kind of user-generated content could be shown to their social network and social media users (Helberger et al., 2020 B, 12). In this sense, they are intermediaries acting as “information gatekeepers” (Bukovska, 2020, 28) who can decide whether particular opinions, thoughts and information should be allowed to pass the gate to meet the users or not.

II. AI Content Curation and Moderation as Challenges to Freedom of Expression to Promote Gender Equality on Social Media

AI content curation and moderation operate social networking and social media platforms have several significant impacts on all types of online expression, ranging from political and social to artistic and commercial expression. It is unavoidable for gender equality, as an issue of political and social expression on the Internet, to be affected by AI technologies to a great extent. In other words, AI content curation and moderation could pose challenges to online freedom of expression in general and to the expression concerning gender equality for women in particular.

Firstly, the AI content curation systems potentially limit the users’ access to a wide range of ideas, opinion and information, since the information appearing on the users’ feed is normally selected and recommended by the AI content curation based on the preferences of individual users. Technically speaking, this means that, if they have different preferences, it is typical that the personalizing information which User A sees on his feed is always different from the personalizing information which User B sees on her feed. As pointed out in the previous section, in democratic societies, good policies can be only achieved through intelligent voting which is based mainly on the participation in public discussions with various ideas and opinions, including pros and cons on a particular social or political issue, are debated. In principle, social media are expected to be fora for public discussions. Nonetheless, the way in which the AI content curation operates could prevent people, especially social media users, from being exposed to a variety of views and information since they would be only fed with the AI-recommended views and information that are compatible with their preferences. As a matter of principle, with the limitation of access to a wide range of opinions, it is hard to hope for the intelligent vote.

Secondly, the AI content curation systems could trap the users in echo chambers (Bukovska, 2020, 60). Because of the AI content curation, the range of different information appearing on feeds is narrowed down to a small pool of AI-chosen information and content. This could create “environments in which the opinion, political leaning, or belief of users about a topic gets reinforced due to repeated interactions with peers or sources having similar tendencies and attitudes” (Cinelli et al., 2020, 1).

As a result, in the social media echo chambers, the users would see or read only the thoughts and the views that they agree with or that do not challenge their pre-existing ideas and opinion. Thus, the users are likely to avoid or overlook ideas and opinions which differ from theirs, and worse, are prone to “more extreme positions” (Cinelli et al., 2020, 1) on particular social or political issues. As the echo chamber effect on social media limits ideas, opinions and information available to individual users, it is an obstacle to freedom of expression (freedom to access to various ideas, opinions and information) and democracy in the same manner as the first challenge (argued above) does.

Thirdly, the AI content curation hinders the organization of online campaigns and social movements. To organize online campaigns and social movements, it is requisite for the organizers or the campaigners to gain sufficient supports from a number of social media users to build a collective effort with a strength to launch a campaign or a social movement. Therefore, it is essential for the messages from the organizers or the campaigners to reach other users as many as possible. Nonetheless, as argued by Natali Helberger et al. (2020 B, 13), “[if] algorithmic [personalization] is taken to the extreme, combining algorithmic gatekeeping with AI-driven content production,” it is possible that a piece of information relating to a social matter (such as a news article) might, in the end, reach just one and only person who is interested in that particular issue, whereas other people might receive different pieces of information (which are chosen and fed to them by AI). In this regard, traditional media (for example, televisions and newspapers) appear to be a better and more effective channel to deliver messages to a wider audience in comparison to social media, because the information imparted from the traditional media could reach wider audiences regardless of their preference, since traditional media-disseminated information cannot be customized for individual viewers or readers. As a consequence, “a common communication space where public agendas are formed” (Helberger et al.,2020, 13) could be created, making people have a sense of being a part of a collective effort and consolidate their powers. This could allow campaigns or social movements to develop and happen eventually. However, as the AI content curation, in effect, fragments users by means of delivering different customized information to different individual users. Put differently, different users would always receive different messages about different matters. Given this, it is difficult to expect the common communication space where the users who have similar thoughts and opinions or share the same social values can (virtually) gather to be created. As a consequence, without the feeling of being a part of a collective effort and the consolidated powers, the initiation and the progression of online campaigns and social movements would seem unlikely to happen.

Last, but not least, is the issue of AI content moderation which has a task inter alia to monitor and remove so-called ‘problematic’ content. Technically speaking, it is, in fact, the Internet censorship powered by AI which is employed by private companies. There are some points to be noted here. First of all, as stated in the previous section, the right to freedom of expression is the legal claim to protect people’s ability to hold and express their ideas and opinions against states’ (arbitrary) censorship (Zeno-Zencovich, 2008, 1). However, in the case of AI content moderation, they are the private companies, not the states, which implement the Internet censorship. Disappointingly, the cruel truth is that the constitutional guarantee of the right to freedom of expression does not safeguard the users from the censorship power wielded by the tech companies. Secondly, the removal (censorship) of content is determined by AI algorithms. However, in most cases, algorithms developed by the tech companies are treated as their trade secrets, for example, the ‘PageRank’ algorithm of Google (Schwartz, 2013, 651). It is also the case for social networking and social media platforms. This could raise a transparency issue, as the users will never know how the algorithm in question was designed and how it operates. This could mean that the users’ right to freedom of expression on the Internet as both speakers and audiences is curtailed by AI moderation ‘black boxes’. The non-transparent designs, the invisible deployment and implementation of AI moderation are obviously at odds with a fundamental principle of freedom of expression which requires censorship to be scrutinized by relevant government agencies, the users, as well as the general public through a legal process. The last point to be made here is that it is always possible for the public authorities to put pressure on the tech companies with regard to illegal content (Bukovska, 2020, 19-20); and as a result, to keep themselves safe from troubles, the tech companies might design AI moderation algorithms to have a sweeping performance which would filter out and remove not only undeniably unlawful information, but also information which has not been decided by a court to be illegal yet. Worse, on the slippery slope, AI moderation algorithms might go further to censor even online expression which has social or political values, but may offend the states. Lastly, due to the possible bias against minority groups embedded in the datasets that are used to train AI, content concerning the information and the expression of the minority groups could be filtered out or taken down (Bukovska, 2020, 58).

As far as the issue of gender equality for women is concerned, the AI content curation and the social media echo chamber effect which it creates could prevent people from having an access to diverse outlooks and information that are relevant to gender equality. As a consequence, they would lack opportunities to have sufficient information necessary for participating in the public discussions relevant to this matter. Moreover, women and gender equality supporters in particular would lack chances to learn from their peers in different countries about how to bring changes in social policies in favor of women’s rights in their countries. Likewise, without a full access to information and perspectives pertinent to gender equality, it is not easy for people to vote intelligently which could pave the way for better polices on gender equality which could redress discrimination and unfair treatments against women. Furthermore, as argued above, the AI content curation system (at its extreme level) could deter social media users who enshrine the value of gender equality or those who advocate the ideas that women deserve fair treatments equally to men from uniting and launching campaigns or social movements to call for gender equality. Nevertheless, fortunately, the AI is not that extreme yet at the moment. The success of the #Metoo movement serve as clear evidence. Yet, nothing can guarantee that the situation which Helberger et al. (2020 B, 13) has horrifyingly envisaged will never happen in the future.

With respect to AI content moderation, online expression to promote gender equality for women could be negatively impacted in several ways. Firstly, as the online content moderation is implemented by private companies, the constitutional protection of the right to freedom of expression is inapplicable to the expression to advocate gender equality on social networking and social media platforms. Secondly, due to the protection of AI algorithms under the principle of trade secrets, it is very unlikely or even impossible for gender equality supporters to know whether the AI content moderation systems deal with their online expression properly in accordance with the freedom of expression principles. Even though there may be suspicious cases, it is unpromising for them and the public to have an opportunity to inspect the algorithm in question. Thirdly, because of the secrecy of the AI content moderation systems, people who champion women’s rights and gender equality cannot be ensured that the systems censor only criminal or harmful content, not their messages which are absolutely lawful but are disliked by the states. Lastly, as stated above, information and expression of minority groups of women which call for gender equality, such as lesbians, female sex workers and etc., could possibly be filtered out or removed due to the bias datasets which are used to train AI tools.

As discussed above, the operation of AI technologies behind social networking and social media services inevitably has several detrimental effects on freedom of expression on social media in relation to the promotion of gender equality for women. However, without an attempt to overcome them, not only would the right to freedom of expression be undermined, but also the achievement of gender equality for women would become a downhearted goal. The approach to deal with this problem will be proposed in the following section.

The Two-tiered Approach: Policies and Law

As examined in the previous section, AI technologies which the tech companies utilize to curate and moderate content on their social networking and social media services could, in effect, curtail online freedom of expression and attempts to promote gender equality. This section aims to provide a recommendation to mitigate the negative effects of AI content curation and moderation on online freedom of expression. The approach is two-tiered. The first tier is comprised of policies focusing on the tech companies and their usage of AI technologies. For the second tier, the implementation of these policies should be supported by legal measures. Put differently, the policies should be set to aim for the guideline of good practices which have freedom of expression to promote gender equality as their core; the law should be designed to support the implementation of the policies.

I. The Policies

To make proper policies on the negative impacts of AI content curation and moderation on online freedom of expression to promote gender equality, the helpful starting point is the Council of Europe’s document. It is the output of Conference of Ministers responsible for Media and Information Society “Artificial Intelligence – Intelligent Politics: Challenges and Opportunities for Media and Democracy” which was held online between June 10-11, 2021. The document is the “Final Declaration: Resolution on Freedom of Expression and Digital Technologies …” (hereinafter “the Final Declaration document”) which was adopted on June 11, 2021 (Council of Europe, 2021). It is worth noting that although it addresses the impacts of AI technologies on the right to freedom of expression in general and does not mention about gender equality in particular, its proposed policies could be applicable to the issue of AI and online freedom of expression to promote gender equality to a great extent.

Certain key issues in relation to online freedom of expression which are stated in the Final Declaration document could be summarized as follows: (Council of Europe, 2021, 2-6)

The right to freedom of expression as guaranteed by European Convention of Human Rights (ECHR) is recognized as a fundamental part of democracy and thus needs to be protected. It is important to safeguard this right against undue interference since it allows people to partake in democratic activities and make informed choices.

The application of AI technologies to modern communications platforms (the Internet) has several profound impacts on the right to freedom of expression, especially the access to and the distribution of ideas, opinions and information.

The social network and social media companies have increasingly played the role in selecting and editing information and content available to their users. The personalized feeding of information and content in accordance with preferences of individual users could lead to fragmentation and division of people in a society and, in turn, have undesirable impacts on the development and maintenance of the unity of the society and democracy as a whole.

While algorithms which drive AI tools are helpful to detect and identify unlawful and harmful information automatically, human supervision is still necessary to ensure that undue restrictions on freedom of expression will not occur. (However, the issue regarding labor conditions requires further discussions.)

The existing operations of AI content moderation may not always comply with legality, legitimacy, and proportionality norms of the ECHR. Therefore, to prevent unintended exceeding restrictions on freedom of expression caused by undue takedown, bias and the lack of transparency, the social network and social media companies should provide complaint channel for the users.

The standards regarding the proper operation of AI tools on social networking and social media platforms and its transparency should be developed by independent research and the dialogues between stakeholders, namely civil society, service providers, social media companies, individual users and the media.

Due to the possible bias in the datasets which are used to train AI, ideas, thoughts and opinions expressed by minority groups might not be able to reach the wide audiences.

The limited exposure to diverse ideas, opinions and information due to the AI-selected content in according to the past and existing preferences of individual users could have negative influences on non-interfering self-development and the liberty to form opinions of the users.

Given all issues above, some important recommendations of Final Declaration document are as follows: (Council of Europe, 2021, 6-9)

  • The right to freedom of expression should be respected and in line with the relevant caselaw of the European Court of Human Rights.

  • There should be binding regulation, where appropriate, to provide legal remedies and independent supervisory bodies to ensure that the proper utilization of AI content moderation to prevent exceeding restrictions on the right to freedom of expression.

  • The designs, the development and the implementation of the AI tools should consider ‘human rights by design’ and the risk prevention and mitigation measures.

  • There should be a guidance regarding responsibility, transparency and accountability of the use of AI tools.

  • The attention should be given to marginalized groups to ensure that they are not excluded and can exercise the right to freedom of expression, especially the access to a full range of information.

  • There should be the system of human oversight to ensure that the operations of AI are consistent with the right to freedom of expression principles.

  • There should be reviews, consultations with all relevant stakeholders and reports on measures which will have been taken to implement the resolution in the Final Declaration documents on a regular basis.

As far as freedom of expression on social media in relation to gender equality is concerned, it could be said that all issues addressed in the Final Declaration document are consistent with what discussed in the previous two sections. Thus, the recommendations in this document could serve as a foundation to develop policies on the promotion of gender equality on social media to a great extent. First and foremost, the principles of right to freedom of expression and the goal to achieve gender equality should be on the top of the policy hierarchy.

Secondly, the tech companies should include the concepts of the right to freedom of expression and gender equality into the designs of their AI content curation and moderation tools to ensure that all people, including marginalized or minority groups, would not be unacceptably affected by the implementation of their AI tools. It is particularly recommended that judgements from human rights case law should be included in the datasets to train AI tools to become more sophisticated in filtering out and removing illegal and harmful content without affecting lawful expression regarding gender equality. However, as the human rights courts in different countries give different levels of freedom of expression protection, it is recommended that there should be more research on the question of what the acceptable level of freedom of expression protection should be by examining and comparing judgements pertinent to freedom of expression from as many human rights courts in different countries as possible. The objective of this research is to develop datasets to input in the training processes of AI. Although this research project seems to be over-ambitious, it is not impossible if both international organizations and governments of different countries collaborate. Thirdly, the tech companies should develop risk prevention and mitigation measures, as well as providing complaint channels in the case that their AI tools damage or violate the right to freedom of expression of gender equality advocates. Fourthly, the automated operation of AI content curation and moderation should be under human supervision. (However, how to achieve good practices regarding human workforce needs to be discussed further by relevant actors.) Fifthly, the tech industry should establish an independent self-regulatory body. Its main duties are to set a code of conducts and to oversee whether individual companies comply with the code of conducts. The code of conducts should aim for minimizing the adverse effects on the users’ freedom of expression caused by AI content curation and moderation, i.e., the undue limitation of access to and expression of diverse ideas and opinions about gender equality, the echo chamber effect which negatively impacts the promotion of gender equality, the hinderance of online campaigns and social movements to call for better changes about gender equality and exceeding and inappropriate taking down of information concerning gender equality, the exclusion of the voices of the minority groups. Sixthly, AI content curation and moderation should be employed and deployed with responsibilities. The social companies should be held accountable for the use of AI tools. More importantly, with regard to transparency, when damage or violation of the right to freedom of expression occurs to those who champion gender equality, the company at issue should disclose how its algorithm works at least just only the part which caused damage or violation. However, in the case which the company rejects to do so, the injured party should be allowed to seek justice from a court, asking the court to order the disclosure of the algorithm in question. Lastly, the self-regulatory body and other stakeholder, such as the public sector, the tech industry, the social media users and the relevant civil societies and interest groups should jointly conduct reviews of the overall operations of the IA content curation and moderation and arrange consultation meetings to address the problems and concerns and to find solutions on a regulatory basis. The information regarding the reviews and the meeting should be made publicly available.

II. Law

All of the policies recommended above should be on a voluntary basis. However, to put them into a practical effect, they must be fostered by a legal measure. The legal measure is this context does not mean criminal law or punitive law which aims to punish if someone fails to comply with it. On the contrary, it should be the law which is designed to facilitate the co-operation between all stakeholders (such as the law to encourage the establishment of the self-regulatory to set the code of conduct and oversee the use of AI tools as stated above). Furthermore, it should be designed to be persuasive enough to attract the relevant actors, especially the tech companies to participate and collaborate with other interest groups with the primary goal to strike a proper balance between profit-making and the respect of the right to freedom of expression in relation to the promotion of gender equality on the Internet. The law which gives certain benefits to the tech companies, such as cooperate tax reduction, could be an example of the legal measure proposed in this sub-section.


Millions of women around the world have underprivileged social status, limited opportunities of their lives and experience discrimination and unfair treatment because of gender inequality in their societies. However, the Internet and social networking and social media platforms have provided them not only a very powerful channel to voice their views and thoughts, but also venues for them to (virtually) gather to make powerful calls to the governments for social and political changes through online campaigns and social movements. They exercise the right to freedom of expression, which is one of the fundamental and vital elements of democracy, on social networking and social media platforms to improve the situation of gender inequality. Nevertheless, the tech companies which own social networking and social media services employ and deploy AI content curation and moderation for their commercial purposes. Inevitably, the operations of the AI tools in this manner have negative effects on freedom of expression to promote gender equality, since views and thoughts could be restricted by the AI tools based on their algorithms.

This chapter have examined and discussed these issues. In its final section, it has proposed that this problem could be tackled by the two-tiered approach. At the heart of the two-tiered approach are the policies which give importance to freedom of expression in relation to the promotion of gender equality and the implementation of the policies that is fostered by legal measures. What is discussed and proposed in chapter could give some ideas to all stakeholders about how to harness the AI technologies to avoid undesirable impacts on freedom of expression to promote of gender equality on social media. However, due to limit space, this chapter cannot delve into the details of certain issues relating to the proposed policies, such as the plausible way to implement human oversight of the AI effectively and the creation of datasets of judgements relating to freedom of expression from human rights courts in different countries. Therefore, further studies are still required.


Anderson, M. (2016). 3. The Hashtag #BlackLivesMatter Emerges: Social Activism on Twitter. Pew Research Center.

Auxier, B. and Anderson, M. (2021). Social Media Use in 2021. Pew Research Center.

Article 19. (2020). Freedom of Expression and Women’s Equality: Ensuring Comprehensive Rights Protection.

Barendt, E. (2007). Freedom of Speech (2nd ed.). Oxford: Oxford University Press

Blossom, J. (2009). Content Nation: Surviving and Thriving as Social Media Changes Our Work, Our Lives, and Our Future. Indianapolis: Wiley Publishing.

Bottero, W. (2005). Stratification: Social division and inequality. London: Routledge

Bukovska, B. (2020). Spotlight on Artificial Intelligence and Freedom of Expression #SAIFE.

Carney, N. (2016). All Lives Matter, but so Does Race: Black Lives Matter and the Evolving Role of Social Media. Humanity & Society, 40(2), pp.180-199.

Cinellia, M., et al. (2021). The Echo Chamber Effect on Social Media. PNAS, 118(9).

Council of Europe. (2021). Final Declaration: Resolution on Freedom of Expression and Digital Technologies, Resolution on the Safety of journalists, Resolution on the Changing Media and Information Environment, Resolution on the Impacts of the COVID-19 Pandemic on Freedom of Expression.

Cram. I. (2002). A Virtue Less Cloistered: Courts, Speech and Constitutions. ‎ Oxford: Hart Publishing.

Data Never Sleep. (n.d.). How much data is generated every minute? [infographic].

Fileborn, B. and Loney-Howe, R. (2019). Introduction: Mapping the Emergence of #MeToo. In Fileborn, B. and Loney-Howe, R. (eds). #MeToo and the Politics of Social Change (pp.1-18). Cham: Palgrave Macmillan.

Henderson, J.J. (2013). The Boundaries of Free Speech in Social Media. In Stewart, D.R. Social Media and the Law: A Guidebook for Communication Students and Professionals (pp.1-22). London: Routledge

Helberger, N., et al. (2020 A). A Freedom of Expression Perspective on AI in the Media – With a Special Focus on Editorial Decision Making on Social Media Platforms and In the News Media. European Journal of Law and Technology, 11(3).

Helberger, N., Eskens, S., van Drunen, M., Bastian, M. and Moeller, J. (2020 B). Implications of AI-Driven Tools in the Media for Freedom of Expression.

Khomami, N. (2017, October 20). #MeToo: How a Hashtag Became a Rallying Cry Against Sexual Harassment. The Guardian.

Lively, J. (1975). Democracy. Oxford: Basil Blackwell

Loiseau, E. and Keiko, N. (2015). Can Social Media Effectively Include Women’s Voices in Decision-Making Processes?.

Lorber, J. (2010). Gender Inequality: Feminist Theories and Politics. New York: Oxford University Press.

Llanso, E., et al. (2020). Artificial Intelligence, Content Moderation, and Freedom of Expression.

MacKinnon, C.A. (2020). Global #MeToo. In Noel, A.M. and Oppenheimer D.B. (eds.), The Global #MeToo Movement (pp.1-15). Washington DC: Full Court Press

Marr, B. (n.d.). How Much Data Do We Create Every Day? The Mind-Blowing Stats Everyone Should Read.

Mcmillan, S.J. and Morrison, M. (2006). Coming of age with the Internet: A Qualitative Exploration of How the Internet Has Become an Integral Part of Young People’s Lives. New Media & Society, 8(1), pp.73-95.

Milano, A. [@Alyssa_Milano]. (2017, Oct 16). If you’ve been sexually harassed or assaulted write ‘me too’ as a reply to this tweet. [Tweet].Twitter

Muller, M.G. and Hubner, C. (2014). How Facebook Facilitated the Jasmine Revolution: Conceptualizing the Functions of Online Social Network Communication. Journal of Social Media Studies, 1(1), pp.17-33.

North, A. (2019, October 4). 7 Positive Changes That Have Come From The #MeToo Movement. Vox.

OHCHR. (n.d.). Enhancing Equality and Countering Discrimination.

Pasquino, G. (2008). Populism and Democracy. In: Albertazzi, D., McDonnell, D. (eds), Twenty-First Century Populism: The Spectre of Western European Democracy (pp.15-29). London: Palgrave Macmillan

Rowland, D. (2005). Free Expression and Defamation. in Klang, M. and Murray A. (eds.), Human Rights in the Digital Age (pp.55-70). London: Glasshouse Press

Ryan, J. (2010). A History of the Internet and the Digital Future. London: Reaktion Books

Saward, M. (2000). Democratic Innovation. In Saward, M. (ed.), Democratic Innovation: Deliberation, Representation and Association (pp.3-13). London: Routledge

Schauer, F. (1982). Free Speech: A Philosophical Enquiry. Cambridge: Cambridge University Press.

Schwartz, A.A. (2013). The Corporate Preference for Trade Secret. Ohio State Law Journal, 74(4), pp.623-668.

Statista. (2022). Global Digital Population as of April 2022.

Stepanova, E. (2011). The Role of Information Communication Technologies in the “Arab Spring”: Implications Beyond the Region. PONARS Eurasia Policy Memo, 159.

Stephens, R.T. (2009). Empirical Analysis of Functional Web 2.0 Environments. In Lytras, M.D., Damiani, E. and Ordóñez de Pablos, P. (eds.), Web 2.0: The Business Model (pp.1-20). New York: Springer

Tuerkheimer, D. (2019). Beyond #Metoo. New York University Law Review, 94, pp.1146-1208.

Weale, A. (2007). Democracy (2nd ed.). New York: Palgrave Macmillan.

Yar, M. and Steinmetz, K.F. (2019). Cybercrime and Society (3rd ed). Los Angeles: Sage

Zeno-Zencovich, V. (2008). Freedom of Expression: A Critical and Comparative Analysis. London: Routledge-Cavendish.

No comments here
Why not start the discussion?