Homana S, School Of Excellence In Law, The Tamil Nadu Dr. Ambedkar Law University, Chennai.
ABSTRACT
Artificial intelligence is a combination of opportunities and hazardous for democracy that arises risks in various forms in the society. This research paper discuss the democratic rules in the emerging social media world where the information collected should be based on good faith, ethical and transparent and it also discuss what content should be moderated and discuss some of the advantage and disadvantage. This paper examines some instances of Facebook and how deepfakes affected affected the general election in the India. Further discuss the Misinformation and Deepfakes, AI is biased which censor the voices of the minority voices. This paper discuss the strategies of regulations of AI in social media i.e. transparency, involvement of human in the AI, reviewing of artificial intelligence system regulary and framing of regulatory frameworks. Finally it discuss about the education of artificial intelligence to the citizens, promotion and improvement of independent journalism.
Keywords: Artificial intelligence, democratic rules, transparency, misinformation, deepfakes, human involvement in AI, awareness, journalism.
ARTIFICIAL INTELIGENCE BETWEEN OPPORTUNITIES AND DANGERS FOR DEMOCRACY
The computer network introduction in the world has brought deep changes to our societies, this transition that followed to develop a new and productive applications, such as virtual agents. Artificial intelligence makes sure the efficacious and innovation and it could bring a total change in healthcare and education. On the other hand there arise a risks in ecology, society, and above everything, democracy gets affected.
DEMOCRATIC RULES IN THIS DIGITAL WORLD
Democracy depends on the rules even in the digital age and the rules must be applied .
Social responsibility
The informations been collected through the digital networks must focus particularly on good faith. Personal data of the citizens must be used only for the improvement of public services like Health care and education, facilities for disabled and this artificial intelligence shouldn’t be used for the manipulation of the people or control the people. The tech companies plays a major role in this era, the tech companies are encouraged to revise their business models, and it could be better if system is more ethical and transparent.
Decentralization
If the entire information of the society is in fewer hands then it would be danger to democracy , clearly it represents the danger to democracy. So, no entity should have unlimited access to all the data. Decentralization of information is essential to caution against the abuse of powers and centralized control.
CHARACTERISTICS OF AI IN SOCIAL MEDIA AND DEMOCRACY
Artificial intelligence (AI) is quickly becoming the foundation of social media platforms, influencing how users interact with material, communicate with one another, and even develop ideas. AI influences billions of consumers’ digital experiences in a variety of ways, including content control and tailored suggestions. However AI produce various benefits to these social media platforms, it even brings forth dangers that endanger democracy. We will examine at the two nature of AI in social media—its potential pros and cons—through important instances and legal consequences.
AI for content moderation and personalization
Artificial Intelligence (AI) mechanism are created to cut, censor, blur and recommend videos or messages based on the activities of the users. Social media Platforms such as Facebook, YouTube, and Twitter given employment for AI to monitor user connections, evaluate users interest, and take care to what material should displayed in their explore page or (feeds). By checking larger quantity of data collected, Artificial Intelligence(AI) guide to adapt each user’s experience, encouraging by producing material that is likely to be fascinating based on precedingconnections[1].
Nevertheless, AI’s involvement in content instruction and modification as per the interest of the users increase significant ethical issues. AI’s main primary employment is to make the most of user interaction, yet this objective commonly encourages unbelievable, fervently charged, or argumentative material. This forms of material leads to prompt solid excitements, which increases engagement numbers. The challenges is a platform environment that often brings forth the amplification over keen, balanced conversation. The procedures that rule social media platforms lead to emphasize recognized ideas, resulting in “echo chambers” that underwrite to political separation and citizens thoughts disintegrated[2].
However Artificial Intelligence (AI) can efficiently eradicate hazardous material, like hate speech or deep fake or graphic content , it does have limitations. For instances, AI cannot entirely understand the background of a message, that leads in misconceptions that would obstruct productive argument or silence specific voices.
Facebook’s AI system, for instance
Facebook’s content restraint system, which employs machine learning algorithms to recognize and reject hate speech, is a well-known example of how Artificial Intelligence is affecting social media platforms. Facebook privileges that its AI expertise has effectively cut hate speech on the network system by about half in last two years. This is a noteworthy achievement in the fight against harmful social media data[3].
There are various noteworthy disadvantages, though. Facebook’s AI technology rarely made a mistake by reporting content that did not breach its community rules. Racial justice activists’ posts, for instance, were incorrectly accused on inspiring violence. This brought the problems of unnecessary censoring or bluring in addition to highlighting AI’s limitations in decoding complex messages. It is very clear the AI algorithms can’t sense or unable to understand the context and the usage would be unreasonably stifle the significant messages or arguments, specifically which refers to the social and political problems[4].
This could be a serious problem for democracy. AI structures has the possible function to silence the important voices and stop the free discussions of ideas, which is very much needed for democratic address, if the regulation of AI are not transparent or well standardized. Further, the strict AI filtering might hamper the users capacity to share the difference of opinions, restriction of free speech, and defeat social movements.
Google v. Oracle, United States, 2021[5]
This case google v. oracle case has brought front a interesting lawful content for the debate over artificial intelligence’s usage in social media platform and digital platforms. This case deeply discussed the usage of artificial intelligence, data and intellectual property as well as obligations, even though it had nothing related to discuss with social media materials. The U.S. Supreme court ruled in favour of fair usage, holding that Google’s use of Java code in android did not interrupt oracle’s copyrights. Assuming the boundaries or territories if intellectual property law in relationship to Artificial Intelligence and software use was a major conclusion of the case.
Accountability with legal idea is the very most significant lesson that to be learned from this case for social media sites. In the content of social media platform, this case determines how important AI’s transparency and responsibility are. To the knowledge, without reasonable regulation, this AI mechanism have the potential to be employed in such a way that it violates users rights or change the public opinion, this might have a significant negative influence on democracy[6].
Example 2: Election Deepfakes in India in 2024
The use of AI generated deepfake videos to influence voters during the 2024 Indian elections was one of the most relatable uses of Artificial Intelligence in recent elections. Deepfakes are Artificial Intelligence generated materials, specifically videos, which may change the person’s words, phrases or appearances to give the impression that they are saying or doing something they haven’t. by making politicians seem to say things that may defame their reputation or sway voter’s opinions, these AI generated deep fake videos are utilized to fabricate stories. The negative aspects of AI is illustrated in this example. AI may be abused for evil intent, even though if AI is utilized for good reasons, such as boosting user engagement or helping political campaigns. Deepfakes were used as a misinformation technique in this instance, producing a fascinating world that affected voting behaviour. Voters may conclude their judgements on inaccurate or distorted information; hence the production of such false videos has the obvious capacity to disruption of democratic process.
In this digital era trusting the information that influences political opinion crucial in democracies. Un recognised AI technologies, such as deepfakes, have the capacity to undermine public confidence or belief in election procedures and influence public opinion in ways that affect free and fair elections. As a result, artificial intelligence encourages participation and support democracy, it can also be used as a weapon to sabotage[7].
CHALLENGES POSED BY AI IN SOCIAL MEDIA
The social media plat form functions been rapidly transforming through Artificial Intelligence, but it even produces various important challenges for democracy. These challenges could demoralize free speech, spread misinformation, support biases, and develop an environment that can biasedly favour certain groups and individuals. Let’s dive into the particular challenges posted by AI in social media.
Misinformation and Deepfakes[8]
Deepfake generation videos and automated content creation are rapidly increasing. Deepfake audios, videos and images make the users into believing false narratives, which influence public opinion and influence political social decisions.
Example: Rahul Gandhi’s Deepfake (India)
In India, during the 2024 elections, AI Generated deepfake video was showed politician Rahul Gandhi resigning from his position, an event that never happened. This influenced content was made to mislead the public and damage the politicians fame in the society. This AI ability create a realistic videos and voters decision on manipulated or fabricated information, rather than a factual content
Case Law: United States v. Google (2018, US)[9]
This case focused on the obligations of tech companies, including their mechanism in spreading misinformation. This case brought the regulation measures that gives social media platforms responsibility for the content they promote. This case been a guide for companies to legally require to manage the spread of misinformation, particularly when AI systems purposefully increase injurious or ambiguous content.
Algorithmic Bias
AI structure are not unbiased they are influenced by the data they are trained on. This means that if the training data duplicates human biases. The AI can inadvertently continue those biases.
For example, African American Vernacular English was been misinterpreted by AI systems, leading to a wrong positives where honesty, non-offensive posts are labelled as unfortunate. Likewise, expressions used by political party campaigners or marginal groups to highlight communal issues could be wrongly flagged, misleading significant discussions.
Case Law: Brown v. Board of education (1954,US)[10]
This case is indirectly depend on AI , for honesty and without discrimination this case been set a precedent, encouraging that AI shouldn’t continue discrimination.Ai algorithm should be fair without discrimination. This bias should be addressed to prevent the silencing of minority voices[11].
Lacking of transparency
One of the important disadvantage of AI in social media platforms is the lack of transparency in how the system operates and how content is flagged and removed. AI driven system often work behind the scenes, flagging contents on certain criteria or regulations, but the users are rarely informed about why their content was removed or why certain contents in their feed.
YouTube’s Content Control
Even you tube use Artificial intelligence to robotically flag videos for elimination if the contents violate the community instructions or guidelines. But still the algorithm might not always make the correct judgment, and human controllers often come forward to review the flagged content. Here the issue is the users are rarely given a clear explanation of why the content was flagged or why it removed, which causes distrust among citizen. AI can’t differentiate the legalized, non offence content and manipulating or inappropriate content.
This lagging of transparency make the citizen to decrease the trust in the platform and the AI is blamed of biased moderation.
Case Law: The Digital Services Act ( EU, 2020)
The Digital Services Act in the European Union primary goal is to address this transparency problem by demanding digital platforms to be more reasonable for their content controlling practices. Under the Digital Services Act, digital platforms are obligatory to explain their content controlling process unambiguously and to permit the users to appeal decisions about content moderation. This is the essential regulation for fairer and transparent system for content flagging.
Lacking of clear guideline regulations on how the AI is making content moderation, users feel difficulty to argue for the unfair flagging of content. This even lead to monopolization like only some contents are feeded and other contents are moderated or ignored.
REAL- TIME INSTANCES OF AI’S IMPACT ON DEMOCRACY
India’s 2024 Election
In the 2024 general elections in India, voter engagement role was played as a centrally by AI. AI system monitored large amount of data about the voters through the social media activity, interests, and last voting behaviour, to address the précised message.
But, this monitorization brought risks. AI was keep used to make and distribute deepfake videos, and fake false stories. This contents made to mislead voters. Which made the citizens face difficulty to distinguish the fake and real videos.
Impact on Democracy: This AI generated content degraded the democratic process by spreading information, making unreasonable narratives, and distracting the voters before understanding the key problem. Voters are manipulated by the false information and bring confusion and the voting behaviour is influenced .
Slovakia’s 2023 Election
In 2023’s slovakia’s election, Artificial Intelligence was again used to customise messages and target voters. The dishonoured involvement of a AI generated deepfake video that misleadingly showed a applicant manipulating election result, claiming that they had equipped the voting counts.
This video rapidly gone viral, this made a ambiguation about the trustworthiness of the election process. That video made the voters believe it and temporary dip in public trust in the election system. This AI generated deepfake video undermined the fame of the candidate , it even also affected the public faith in the democratic process.
Impact on Democracy: This example highlights how AI-generated content, when used maliciously, can severely damage the integrity of the electoral process and the trust citizens place in democratic institutions. When AI is used to manipulate perceptions of democracy, it poses a direct threat to the stability of democratic systems.
Case Law:Cambridge Analytica Scandal(2018,UK/US)
This Cambridge Analytics Scandal is one among the unpopular instance of Artificial Intelligence mis application in modern election. In 2018, it has recognised that Cambridge Analytica, a political consulting firm, had used Artificial Intelligence methods to produce and analyse the personal details of millions of Facebook users without the prior consent. Then the firm used the collected data to develop highly aimed political advertisements targeted at influencing the 2016 U.S. presidential election and the referendum in the U.K.
With the help of AI system, this company was able to craft advertisement tailored to individual voters based on their personality traits, psychological displays, and last behaviour. These advertisements were made to manipulate voters, encouraging divisions and strengthening quarrelsome issues. The advertisements mainly targeted on exploiting voter’s fears, biases and emotions specifically topic like immigration and national security.
Legal impact:This case brought the risk of statistic data manipulation and AI system drive the voter influence. It increased the serious worries about the ethics of using personal data for political gain, specifically in shadows of weak regulations governing how data is used in the political campaigns. This scandal encouraged widespread calls for tighter data protection laws and more transparent in political advertising.
Impact on the democracy:This case confirmed that AI not only encourage voters but even manipulate and betray them on a massive scale. The use of AI for these matters has potential to demoralize free and fair elections, this leads to question the fairness of the political system.
STRATEGIES FOR REGULATING AI IN SOCIAL MEDIA
Transparency and accountability: if the contents are moderated or removed by the Artificial Intelligence systems then the tech companies should disclose the reason, this helps the users to understand moderation decisions. the transparency is important to prevent biased decisions or arbitrariness and develop trust on the platforms.
The Digital Services Act (EU, 2020) directs that the platforms explain their content moderation process, promoting responsibility and ambiguation.
Human in the loop Moderation: whenever AI decides to flag or remove the content which is harmful, the AI’S decision should be reviewed by the humans particularly for sensitive contents. Artificial intelligence is not perfect so Human judgement is necessary to avoid misinformation or restriction of genuine content.
The twitter’s moderation system uses both Artificial Intelligence and human moderators to make a accurate decision making, avoiding errors made solely by procedures[12].
Bias Audits and Ethical guiding principles: Artificial intelligence must undergo continuous regular reviews to find out and make correction in the biases, ensuring fairness and equality in content moderation. Bias in Artificial Intelligence structure can lead to discriminatory practices and defeating of certain viewpoints or groups.
[13]IBM’s Artificial Intelligence Ethical guidelines highlight the fairness, transparency and responsibility in Artificial intelligence development, the regulations that social media companies must adapt the sureness of ethical practices.
Regulatory framework: government agencies must create laws that legalise artificial intelligence in political processes, such as giving disclaimers for the content created by the deepfake videos created by artificial intelligence. This government framework ensures artificial intelligence is used ethically and safely, particularly in specific areas like elections. The Artificial Intelligence Act (EU,2021) Set rules for high risk artificial systems, includes those used in social media content moderation, to avoid damage and guarantee safer use.
THE ROLE OF SOCIETY AND PUBLIC AWARENESS
Digital literacy: citizens must be educated on how artificial intelligence works and educate how to identify the misinformation, ensuring they can critically evaluate online content. The well educated about the digital contents can resist the manipulation and false information, by strengthening information. Finland’s National Digital Literacy Program educate students to find the fake news, authorizing people’s to engage with media more critically.
Community based oversight: various groups should be involved in making decisions regarding content moderation, ensuring no group is unfairly censored. It ensures that control reflects a wide range of perceptions, making platforms more democratic. For instance Reddit Community Moderation allowing a volunteer moderators from diverse communities to review and manage content by promoting more inclusive content.
Support for Independent Journalism: social media platform i.e. digital platforms must provide supporting trust worthy, individual media that provides truthful, impartial information. Independent journalism is decisive in countering false information and promoting a knowledgeable constituency. This could promote highly qualified journalism over scandalous or fake content, which can encourage strong public treatise and making decisions informed. By merging regulatory frameworks, public responsiveness and social engagement, we guarantee AI is used to support, rather than demoralize, democratic values.
CONCLUSION
Artificial Intelligence has the power to support democracy by cultivating voter engagement and contact the information. Nevertheless, it even brings hazards such as false information, unfairness and restriction. To confirm artificial intelligence serves democracy, we want sturdy guidelines, transparency and active public participation. Governments, social media platforms and companies and users must cooperate to ensure artificial intelligence assists the public good and without harming.
References
[1] “How Artificial Intelligence Powers Personalized Content Recommendations,” Harvard Business Review.
[2] “The Echo Chamber and Political Polarization,” American Political Science Review.
[3] “Facebook’s Hate Speech Moderation: A Success or a Flaw?” TechCrunch
[4] “Missteps in AI Content Moderation: A Closer Look,” The Verge
[5] “Google v. Oracle: Supreme Court Decision and Its Impact,” SCOTUS Blog
[6] Accountability In Algorithmic Copyright Enforcement Maayan Perel & Niva Elkin-Koren Cite As: 19 Stan.Tech. L.Rev. 473 (2016)
[7] https://www.bbc.com/news/world-asia-india-68918330
[8] The Hindu. (2024). Deepfake video of Rahul Gandhi resigning creates confusion https://www.thehindu.com/elections/lok-sabha/congress-using-fear-rumours-and-now-deepfakes-for-its-propaganda-anurag-thakur/article68127918.ece
[9] United States Supreme Court. (2018). United States v. Google, Inc., 139 S. Ct. 2632
[10] United States Supreme Court. (1954). Brown v. Board of Education, 347 U.S. 483 https://www.supremecourt.gov/visiting/activities/pdf/portraithunt.pdf
[11] Electronic Frontier Foundation. (2023). AI and Content Moderation: The Struggles of Marginalized Voices
