Author- Ashwin Pandey

“Freedom of speech implies the world isn’t defined. It is meaningful when people are allowed to see the world their way”- Ai Weiwei.

Artificial Intelligence is a term that is often used, and yet a lot of us do not have a proper understanding of what it is. Due to this lack of understanding we fail to realize the potential that AI possesses. In a Report by the Special Rapporteur before the United Nations General Assembly, Artificial intelligence was referred to as our ever-increasing dependence on automated decision making. AI is “a constellation of processes” which help enable computers in performing tasks and solving problems that would otherwise have to be done by humans.[1] It plays a pivotal role in shaping how information is accessed by people and how they are able to interact with devices.[2] However, even though Artificial Intelligence does aid in the rapid advancement of technology and has immense potential for enriching and revolutionizing our society, it also poses some threats that must be addressed.

A global human rights organization called article 19, working alongside Privacy International, released a paper called Privacy and freedom of Expression in the age of Artificial Intelligence in which they discussed, in detail, the threats posed towards the freedom of expression by AI.[3] They talk about how AI will have a significant impact on the right to freedom of expression with online intermediaries such as social media using AI to control the information that is made available to users.

The invisible nature of AI, and its ability to analyse the behaviour of the user, can have huge ramifications on the freedom of expression of the user. The user could be forced into self-censorship, or altering their behaviour in public as well as private spheres. With the advent of facial recognition and other such technologies, which provide virtually God-like power to the State, there has been severe hindrance in the freedom of expression as well as the privacy of the individuals in the form of mass surveillance and data collection.[4]

AI poses some critical consequences for the information environment. It can be found in every corner of the internet which has led to many concerns being raised.  

First there is the personalization of content. Social media is the primary way through which common individuals are able to access information along with sharing their own ideas. AI regulates the content which is to be made available to each individual, it uses datasets to create personalized algorithms which curate information on the basis of the user’s browser history, demographics, and other factors.[5] News sites choose the information that needs to be displayed on the basis of the individual rather than important developments. They use AI applications that predict the news patters of a user and their interests,[6] this minimizes the exposure of the user towards diverse views which further risks the reinforcement of inherent biases and prejudice. AI also gives lower priority to content that has lower levels of engagement, so independent as well as user generated content usually gets banished into irrelevance and obscurity, which clearly has a negative effect on the freedom of expression of the creators. This shows the hidden impact that AI has on us daily.

The second concern is with regards to the moderation of content. AI is used by social media outlets to moderate and regulate the content in order to comply with their terms of use. If these terms are violated then the AI is used to either warn the account of possible deactivation, or it removes the same. Companies justify the usage of AI in content moderation by pointing out the sheer volume of illegal content that is posted online which cannot possibly be moderated by humans alone. They further state that AI is better at identifying illegal content than their human counterparts. However, these attempts to moderate content come at a cost to human rights. Content moderation by AI suffers from numerous drawbacks which include taking into account the context behind a post, variation of usages in languages across communities and cultural diversity. AI is often based on datasets that have discriminatory assumptions incorporated into it,[7] and since in many situations there is a low cost attached to over moderation, these posts get struck down. This leaves the vulnerable groups at a particular disadvantage with regards to their content being moderated, and has a negative impact on their freedom of expression.  

There have also been concerns raised with regards to profiling and targeting. AI has incentivized data driven business models that the internet possesses today. The free content that we consume, we pay for in the form of our data.[8] With all of this personal data being gathered over many years, companies use AI to make more precise targeting models. Advertising is not carried out in a generalized manner anymore; it has become extremely pin-pointed. Companies can micro target consumers in order to make them respond to their marketing. This has incentivized the large-scale collection of personal user data which can then be misused to exclude certain sections of society from particular opportunities, etc.[9]

Artificial Intelligence also poses threats to the right to hold opinions.

The right to hold one’s own opinion has been enshrined in Article 19(1) of the ICCPR as well as Article 19 of the UDHR without restrictions. Holding an opinion includes forming and developing them by using reasoning that is free from coercion.[10] Today, everything from our search history, to the emails we send and the documents we store provide a broad outline of the opinions we hold. But when this data is made available to the State as well as private corporations, there is always a chance that they look to interfere in the same.

An even bigger threat is posed to the right to freedom of expression,[11] as we have discussed before, AI is used to moderate content as well as censor the media. There is an increasingly small group of intermediaries who are the gatekeepers to this dissemination of information.[12] The priorities of these bodies may not include hearing a diversity of opinions and they may not promote independent voices in speaking out. Users do not have clarity on the extent of the scope of AI which makes them unable to understand when and how information is disseminated to them.

In India, the NITI Ayog formulated the National Strategy Paper on AI in order to be able to regulate it, with the Government looking towards collaboration with the private players to foster innovation. But it is important to remember that AI operates in a vacuum, and as such it can bring about severe restrictions if the downsides are ignored. The paper concedes that there is a possibility of bias in neutral data. It talks about the black box phenomenon which is the inability to draw a linkage between the input and output due to the algorithms being opaque.[13]

The report goes on to talk about collecting information from platforms in order to employ predictive policing, or the ability to detect a crime before it has happened. A similar system was used in the US with disastrous consequences. The programmers would input corrupted data which would reflect the stereotypes that were held by the police against minorities.[14] A similar thing could happen in India with predictive policing, it could result in targeting minorities and accentuate the prejudice against them. The misuse of AI has already been seen with countries such as China, who have used facial recognition to identify and incarcerate demonstrators, and also used AI towards punishing all the people who held a stance against the Government.[15] This is just one cautionary tale of the possible implications for our democracy if AI were to be left unchecked.

Thus, AI possesses the capability to alter the contours of free speech as we know it, it has obvious drawbacks that have been discussed through the course of the article and there are recorded instances of AI being misused by the states as well as private players[16] for their own gain. The AI policy in India needs to be remodelled keeping all of these concerns in mind. There is a need for comprehensive regulation in order to prevent any adverse effect on the freedom of speech and expression of the people and AI must not be allowed to be weaponized. 

[1] United Nations, General Assembly, Mapping AI’s Impact on Freedom of Expression: report of the special rapporteur, A/73/348 (29 August 2018), available from AI-and-FOE-GA.pdf (

[2] Julia Haas, Freedom of the media and artificial intelligence, The Global Conference for Media Freedom, (Nov. 2020), at p.3-4.

[3] Article 19, Privacy International, Privacy and Freedom of Expression In the Age of Artificial Intelligence, (April 2018).

[4] Supra, Note 1.

[5] World Wide Web Foundation, THE INVISIBLE CURATION OF CONTENT: Facebook’s News Feed and our Information Diets, (April 2018), at p. 14.

[6] Emerging Technology from the arXiv, How Reuters’s Revolutionary AI System Gathers Global News, mit technology review (November 27, 2017). How Reuters’s Revolutionary AI System Gathers Global News | MIT Technology Review

[7] Solon Barocas, Andrew Selbst, Big data’s disparate impact, California Law Review, vol. 104, No. 671 (2016).

[8] Toni M. Massaro, Helen Norton, SIRI-OUSLY? FREE SPEECH RIGHTS AND ARTIFICIAL INTELLIGENCE, Northwestern University Law Review, Vol.110 No. 5, (Mar. 2016), at p. 1177.

[9] Julia Angwin, Ariana Tobin, Facebook (Still) Letting Housing Advertisers Exclude Users by Race, propublica (November 21, 2017). Facebook (Still) Letting Housing Advertisers Exclude Users by Race — ProPublica

[10] Supra, Note 1.

[11] Dr. Natali Helberger, Sarah Eskens, Max van Drunen, Implications of AI-driven tools in the media for freedom of expression, Institute for Information Law (Mar. 2019), at p. 20.

[12] Supra, Note 2.

[13] Rongeet Poddar, The Interface between Artificial Intelligence and Free Speech: Implications for India (Part I), The contemporary law forum (August 25, 2020). The Interface Between Artificial Intelligence And Free Speech: Implications For India (Part I) – THE CONTEMPORARY LAW FORUM (

[14] Rongeet Poddar, The Interface between Artificial Intelligence and Free Speech: Implications for India (Part II), The contemporary law forum (August 25, 2020). The Interface Between Artificial Intelligence And Free Speech: Implications For India (Part II) – THE CONTEMPORARY LAW FORUM (

[15] Ibid.

[16] Balkin, Jack M., Free Speech in the Algorithmic Society: Big Data, Private Governance, and New School Speech Regulation, UC Davis Law Review, Vol. 51, (Sept. 2017), at p. 1173.

Please follow and like us:

By admin

Leave a Reply

Your email address will not be published. Required fields are marked *