AI as a public good

It’s that time of the year when electoral booths become the hotspot. After all, more than half of the world is marching toward elections this year (Over 50 countries have elections this year). With many developments interluding with this event, technology too has found its potential roles and responsibilities. Or have they?

In a recent framework released by the Forum of Information & Democracy, titled “AI as a Public Good: Ensuring Democratic Control of AI in the Information Space”, AI has emerged as more than a tool; it’s a catalyst for unleashing human potential.

As of today, across various information systems, AI is primarily revamping how we create, interact, and access technology. While the curiosity is infectious, its premature deployment has caused a frisson among technological enthusiasts. AI pictures, videos, and other human language systems are causing curiosity, hatred, and fear as well as a growing interest like every technological wave has witnessed. In India alone, the ruling Government’s fresh take on the AI advisory through the ‘India AI Mission’, is aimed at setting up a comprehensive ecosystem. While technology plays a critical function in shaping public opinion, influencing the democratic discourse, and defining the future of the electoral process, the impact of AI produced by both good and malicious actors has to be analyzed. Even the World Economic Forum’s Global Risks Report 2024 refers to AI and social media-driven misinformation and disinformation and highlights the urgency to discuss its impact on elections, political polarisation, and the larger information ecosystem.

India is heading towards the largest-ever election in the world, surpassing the 2019 general elections. Subsequently, it will also be the longest-held general election in the country spread across 44 days. 

The Election Commission of India was established in accordance with the Indian constitution on 25th Jan 1950. The introduction of ICTs in the electoral process within constitutional provisions enriches a commitment to uphold regular, free, and fair elections. While many technological interventions such as; the computerization of the electoral rolls of 620 million voters in 1998, followed by the Elector’s Photo Identity Cards (EPIC) which was first introduced in 1993 to the most recent e-EPIC that brought seamless access to voter-related information under the Digital India scheme, among others have made an entry into this democratic system over the years and brought significant change making the Indian elections one of the most sought after democratic processes. In this context, the intervention of AI will be interesting to explore. 

Citizens v/s consumers

The current narrative about AI focuses on the development and deployment of AI in different industries. From tech to healthcare, to robotic versions of TV hosts to old music surfacing in new versions with a topping of AI, it only normalized its existence. It disengages with the fear it initially brought. While a lot can be explored as consumers of the new technology at hand, our priority remains its impact on the elections.

The prerequisite for every political leader standing at the top is his election campaigning- marketing oneself as the leader of the people, the changes, and amendments one will introduce in favour of the people voting him/her to power. With the influence of AI, this  domain is all set to see the influx of interactive voice responses, and audio and video messages in diverse languages solving the age-old challenge for local leaders to democratize and scale their voter base. This, in itself, is a breakthrough application with the potential to influence the minds of millions of people who will cast their vote. But this intervention has made way for misinformation or its possibility to formulate ethical guardrails and self-regulation within the makers and producers of this new technology.

Another such feature that touches upon diversity is the microtargeting of voters through ads, promotional messages, and calls- based on caste, colour, sexual orientation, gender identity, language, religion, political preferences, property, and other parameters poses significant risks to fundamental rights. The same extends to social media feeds being influenced by algorithms that consider many such filters before throwing content on one’s feed. These risks encompass issues such as manipulation, discrimination, privacy violations, and deception. The FID report says,

“In light of these concerns, AI developers and deployers of AI systems used for ad targeting and delivery based on protected characteristics and special categories of personal data should be held liable for any harm or damage caused by these systems, regardless of the presence or absence of negligence or intentional wrongdoing. Conversely, considering the less intensive invasion of privacy associated with AI systems used for behavioral or contextual targeting, their developers and deployers should be liable for failing to comply with the required risk-mitigation measures, transparency requirements, and duty of care unless evidence proves otherwise, or if it can be demonstrated that they have engaged in wrongful actions, thereby causing harm or damage to another party.”

There is already a concerning lack of deliberation on AI and its implications from the perspective of citizens. Now, as elections close in on us and politicians seek the Midas touch of AI, the million-dollar question falls back upon us, the voters: to think twice.

Can AI be a public good?

The International Monetary Fund defines a public good as “those that are available to all (“nonexcludable”) and that can be enjoyed over and over again by anyone without diminishing the benefits they deliver to others (“nonrival”). The scope of public goods can be local, national, or global.” Now to juxtapose this definition with the possibilities of AI may lead to a mismatch in its current form of design and its hurried deployment is raising ethical considerations, regulations, power dynamics, minority/marginalized rights, privacy, data, etc. However, it can lead to timely interventions in healthcare, education, road safety, and other possible workforces making machine language a part of setting systems that will leapfrog human interventions at a few lousy systems.

AI as a public good’ can throw challenges and solutions.  The FID report says; 

“By framing AI as a public good, we are advocating for a shift in priorities. AI systems must be safe, fair, and reliable if they are to be an innovation that benefits all people and promotes sustainable development. The democratic governance of the information and communication space is a precondition for democracy itself. The responsible use of AI can promote this democratic governance or, on the contrary, mitigate it. The series of measures proposed in this policy framework enables democratic control of AI in the information space.”

In an influential setting like the elections, where citizens are at the core, the confluence of AI warrants an active consultation with the government, citizens, the makers of the technology, the varied systems that are seeking benefits from it, and policymakers. There is an urgent need to set up accountability and redressal mechanisms, that are combed into risk-based approaches of any technological deployment. The ideal approach also will necessarily consider primary rights of citizens such as the right to be informed, to not be discriminated against, and to receive explanations when sought.

Personally, in every such scrutiny or regrouping, the goal has been to address the loose ends and to study the systems better, challenge and verify the compliance to head towards setting up trustworthy systems. Especially, the one that serves democracy.

Shruthi Manjula Mohan

Shruthi is the Program Lead and Opinion Editor at DataLEADS.

Leave a Reply

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