Opinion: Is Artificial Intelligence strengthening or weakening democracy? Reassessing democratic governance in the algorithmic age

[This is an excerpt from an article in The Round Table: The Commonwealth Journal of International Affairs and Policy Studies. Opinions do not reflect the position of the editorial board.]

It is evident that the AI is deeply entrenched in various political organisations and civil institutions across the Commonwealth. Given this scenario it is essential to ask the question, ‘Is Artificial Intelligence Strengthening or Weakening Democracy?’ While AI can be useful for more efficient and effective public service delivery, reducing opacity in political systems, and strengthening public participation, on the other hand, it can also endanger public communication and shared understanding across communities, manipulate opinions, and lead to flawed or ineffective decision-making by citizens (Bauer et al., Citation2025). A democracy relies on public discussion, active participation by citizens and trust in public institutions. AI can be helpful in strengthening these arms of democracy; however, if not regulated properly, AI can also lead to democratic decay.

Since Plato’s Republic, the concept of democracy has evolved from mere elite participation and control to an expectation of active citizen involvement. Political engagement of citizens is now considered to have an instrumental function in the political process, but if unchecked or misled by the authority through the mechanism of ‘polarised pluralism’, it can also lead to totalitarianism (Sartori, Citation1962, p. 77). When a society is extremely polarised, voters may be emotionally overwhelmed, which in turn effects their rationality and compels them to make decision based on extreme ideologies; hyper-participation can be exploited by a totalitarian leader to hijack power in the system. Given the manipulative power of Artificial Intelligence, it has huge potential to give birth to the ‘new age Hitler’. Political activism, governmental responsibility, the freedom of citizens to voice diverse opinions, and a legitimate information system supported by a free and independent press are all essential for the preservation of democracy. For citizens to examine the performance of their government properly, they need to have sufficient and valid information about the system; this, in turn, makes governments accountable (Dahl, Citation1989). In ideal conditions, technology should strengthen these democratic mechanisms, but AI is unlike any other technology!

The vast complexity of AI algorithms makes them very difficult for human understanding, which reduces openness in the information system (Bullock, Citation2019). Without proper regulation, these AI models can use a large amount of data to generate misinformation and spread disinformation. Humans are naturally more attracted to emotionally charged content, be it related to fear, love or violence. If AI models are instructed to generate maximum public engagement, by reviewing the historical pattern of viewers’ engagement, algorithms will automatically focus on highlighting emotionally charged content (Compton, Citation2003). By showing what people already believe, Artificial Intelligence can strengthen these beliefs, leading to a more intolerant society with political divide (Sunstein, Citation2017). AI generated images and videos, such as deepfake content, enhance disinformation and distorted perceptions. Invisible algorithms are being used for content moderation, large scale public surveillance and damaging institutional credibility. Contemporary society is increasingly being governed by the ‘black box’ without any significant accountability (Pasquale, Citation2015). This was starkly illustrated in the Cambridge Analytica scandal, in which the political consulting firm harvested personal data of more than 87 million Facebook users for algorithmic profiling in 2016 US presidential elections. Using this data, Cambridge Analytica identified ‘persuadables’ – undecided voters in key swing states – were bombarded with advertising content and messaging expressly to influence their voter choice; revelations of this manipulation triggered a major push back against Facebook (Hu, Citation2020)

Rather than focusing on whether AI will affect democracy, the focus should be on how AI can be channelled to empower the democratic system. A government’s ability to fulfil public demands legitimises its authority (Easton, Citation1953), and AI has the potential to assist governments in achieving this goal. AI can process huge amounts of data, which offers the opportunity for significant improvement in health care services, the transportation of goods or identifying potential civic unrest stimulated by social media platforms, and even improving government-citizen relations by building interactive platforms. AI-related technologies and big data, when combined, can thus reshape the existing political paradigm, opening up scope for more diffused forms of political participation beyond voting in sporadic elections. It can also help bridge the complexity of understanding democratic processes and institutions by creating a user-friendly interface. If deployed properly, AI can contribute to participatory democracy and effective governance (Helbing et al., Citation2019), leading to the strengthening of democracy.

Tapas Kumar Sahu & D. Purushothaman, Centre for South Asian Studies, Pondicherry University, Puducherry, India.