AI IN GOVERNANCE: MEANING, CHALLENGES AND WAY FORWARD

Copyright infringement not intended

Picture Courtesy:  INDIAN EXPRESS

Context

The Centre introduces 'SabhaSaar', an AI-powered tool, to automate Gram Sabha meeting minutes, starting August 15th in Tripura, signifying AI's growing application in Governance.

Meaning of AI in Governance

AI in Governance means the application of AI technologies to enhance public administration, facilitate service delivery, and inform policy-making, to make government operations more efficient, transparent, and citizen-centric.

Legal Framework for AI in Governance in India

India's approach to AI governance is built upon existing legal structures. Currently, there is no single law dedicated to regulating AI.

Existing Legislation

• Information Technology Act, 2000: Primary law governing the digital space, and its broad definitions include AI systems.
  • Sections 66C (identity theft), 66D (cheating by personation), 66E (privacy violations), and 67A/67B (obscene content) applicable  to AI-generated content like deepfakes and misuse of data.
• Consumer Protection Act, 2019: Protections against harms from AI, particularly through its guidelines against deceptive "dark patterns."
• IT Rules, 2021: Legal obligations on intermediaries to remove prohibited misinformation and deepfakes, ensuring accountability for a safe and trusted internet.
  • Provide a 24-hour timeline for addressing grievances related to morphed or AI-generated images.
  • Aggrieved individuals can appeal to the Grievance Appellate Committee.
• Digital Personal Data Protection (DPDP) Act, 2023: Mandates Data Fiduciaries to safeguard digital personal data and ensures the rights and duties of Data Principals. It provides penalties up to Rs 250 crore for data breaches.
• Bharatiya Nyay Sanhita, 2023 (BNS): Section 111 punishes organized cybercrimes and other sections address cheating and misinformation.

Proposed and Future Legislation: Government plan to replace the IT Act with a new, more comprehensive Digital India Act, to regulate AI-related risks, such as those posed by deepfakes and misinformation.

• In January 2025, a subcommittee under the Principal Scientific Advisor published a report on AI governance for public consultation, a move towards a more formal regulatory framework.

Sector-Specific Regulations: Various sectoral bodies are developing their own AI guidelines.

• Indian Council of Medical Research (ICMR) has released ethical guidelines for AI in healthcare.
• Reserve Bank of India (RBI) is studying AI adoption in the financial sector.
• Department of Defence has outlined a risk-based framework for integrating AI into its operations.

CERT-In: Advisories and guidelines, including safety measures for AI applications, best practices for Generative AI, and supply chain security for emerging technologies.

• Launched the Certified Security Professional in Artificial Intelligence (CSPAI) program to equip cybersecurity professionals with the skills to secure AI systems and address AI-related threats.

Evolution of AI In Governance

Early Stages: As early as the 1990s, India piloted "expert systems"—a basic form of AI—to assist in public administration functions like healthcare and job allocation.  

Strategic Shift (2018 onwards): NITI Aayog released the National Strategy for AI (NSAI), a vision of "#AIforAll," to leverage AI for socio-economic development in key sectors like healthcare, agriculture, and education.  

• Promoting states as a "facilitator" rather than a primary developer, encouraging liberal economic policy, recommending that excessive government investment could "crowd out" private enterprise and that regulation can stifle innovation.

Deepening Focus: Following the NSAI, NITI Aayog published "Responsible AI for All" roadmaps, emphasizing a self-regulatory approach to balance innovation with risks.

• In 2020, the Ministry of Electronics and Information Technology (MeitY) formed expert committees that recommended making datasets more available to promote AI development.

Mission Mode (2024 onwards): Launch of the IndiaAI Mission with an outlay of over ₹10,300 crore, to establish a comprehensive ecosystem with pillars focusing on building compute capacity (over 10,000 GPUs), encouraging innovation, developing indigenous AI models, and ensuring a "Safe & Trusted AI" framework.

What are the ethical Framework and Principles in AI Governance?

Core Principles: The guiding principles for responsible AI governance include:

• Fairness and Bias Mitigation: Ensuring AI systems operate impartially and do not reinforce societal biases.
• Transparency and Explainability: Making AI decision-making processes understandable and open to scrutiny to avoid a "black box" scenario.
• Accountability and Governance: Establishing clear responsibility for the outcomes of AI systems.
• Privacy and Data Protection: Safeguarding personal information and ensuring AI systems adhere to privacy laws like the DPDP Act.
• Safety and Security: Protecting AI systems from unauthorized access and misuse.

Key Initiatives by Government to ensure ethics in AI Governance

• IndiaAI Mission: "Safe and Trusted AI" pillar of this mission is dedicated to creating indigenous frameworks, governance tools, and self-assessment guidelines to ensure ethical AI deployment.
• UNESCO Collaboration: The Ministry of Electronics and Information Technology (MeitY) is collaborating with UNESCO on an AI Readiness Assessment Methodology (RAM), align India's AI policy with UNESCO's Global Recommendation on the Ethics of AI, ensuring a holistic ecosystem built on safety and trust.
• AI Ethical Certification: The Telecommunication Engineering Center, technical arm of the Department of Telecommunications (DoT), is working on an AI Ethical Certification Project to develop tools that ensure fairness in AI models.

Applications of AI in Governance

Enhancing Policy and Decision-Making: AI analyzes vast datasets to inform evidence-based policymaking. NITI Aayog has used an AI tool developed by IIT-Delhi to predict socio-economic trends.

Strengthening Public Service Delivery: Bhashini platform uses AI to offer government services in multiple regional languages, overcoming linguistic barriers.

• DigiLocker uses AI to streamline the verification of documents, making citizen services more efficient.

Improving Law Enforcement: AI is used for predictive policing and real-time crime mapping.

• Delhi Police's AI-driven Facial Recognition System (FRS) assists in solving crimes and finding missing persons, though it raises privacy concerns.
• RBI has developed an AI model named MuleHunter.ai to detect accounts used for financial fraud.

Revolutionizing Healthcare: AI-powered diagnostics are improving healthcare outcomes.

• Niramai, a Bengaluru-based startup, uses machine learning for early-stage breast cancer detection.
• The Ayushman Bharat Digital Mission (ABDM) is promoting an ecosystem for AI-driven health research.

Optimizing Agriculture: AI provides predictive insights to farmers. The 'Kisan e-Mitra' chatbot assists farmers with queries about the PM Kisan Samman Nidhi scheme. A National Pest Surveillance System uses AI to detect crop infestations early.

Enhancing Justice Delivery: To improve judicial efficiency, the Supreme Court has introduced SUPACE (Supreme Court Portal for Assistance in Court's Efficiency), an AI-powered tool to help judges with legal research.  

Smart Cities and Urban Governance: Bengaluru has implemented an AI-driven Adaptive Traffic Control System (ATCS) to manage traffic flow automatically.

Challenges and Concern in AI Governance

Job Displacement: AI-driven automation poses a threat to jobs, particularly in labor-intensive sectors. 

• A World Economic Forum study suggested that AI could displace 75 million jobs in India by 2025.
• A Nasscom study highlights that only 17% of the current workforce possesses the skills needed for AI-era jobs. TCS recently announced layoffs impacting 12,000 employees, citing skill mismatches as a reason

Algorithmic Bias: AI models trained on historically biased data, such as police records reflecting disproportionate arrests of marginalized communities, can amplify discrimination in applications like predictive policing.

• Example: Trinetra platform, used in Uttar Pradesh, has been criticized for marking Dalit-majority villages as "high-risk" areas based on historical data from the discriminatory Criminal Tribes Act.

Privacy and Surveillance: The use of AI-powered surveillance technologies, such as facial recognition by the Delhi Police, raises concerns about mass data collection and potential misuse.

Deepfakes and Misinformation: AI-generated deepfakes pose a credible threat to democracy and public trust. Deepfake cases in India surged by 550% since 2019.

• Fraudsters are leveraging deepfake technology to create convincing fake videos of prominent figures like N. R. Narayana Murthy (Infosys co-founder) and Mukesh Ambani (Reliance Industries Chairman) to promote fraudulent investment schemes.

Digital Divide and Infrastructure Gaps: Benefits of AI are not accessible to all. According to NSSO data, only 24% of rural Indian households have internet access, compared to 66% in urban areas.

• Digital divide, coupled with a lack of robust cloud and computing infrastructure in rural areas, limits the reach of AI-driven governance.
• NITI Aayog estimates that only 22% of Indian firms use AI in any business process.

Lack of a Strong Regulatory Framework: India currently lacks a comprehensive, AI-specific law, which creates legal loopholes and uncertainty regarding accountability, transparency, and liability for AI-induced errors.

Cybersecurity Vulnerabilities: AI systems themselves can become targets of cyberattacks. AI-powered phishing and hacking attempts threaten critical national infrastructure in sectors like banking and defense. 

• Indians lost nearly ₹22,800 crore to cyber scams in 2024, an increase driven by AI (GIREM Report).
• The Indian Cybercrime Coordination Centre (I4C) projects that losses from cybercrime could exceed ₹1.2 lakh crore in 2025 if current trends continue unchecked.
• Number of cybercrime complaints has reached nearly 20 lakh in 2024, up from around 15.6 lakh in 2023 (DataLEADS report).

Way Forward for India to strengthen AI in Governance

Enact a Comprehensive AI Law: Draft a dedicated AI law that balances innovation with regulation, classify AI systems based on risk and establish clear rules for accountability, transparency, and bias mitigation. The proposed Digital India Act is expected to address some of these aspects.

Establish a National AI Regulatory Authority: A specialized body, such as an AI Governance Authority (AIGA), should be created to oversee AI ethics, compliance, and risk assessment.

Boost Indigenous AI Development: Ensure digital sovereignty, reduce dependency on foreign AI infrastructure. The IndiaAI Mission is a significant step towards building domestic supercomputing capacity and sovereign AI models.

Create AI Regulatory Sandboxes: Establish regulatory sandboxes where startups and businesses can test new AI applications in a controlled environment with temporary relaxations on certain regulations.

Promote Explainable AI (XAI): Policies should mandate that AI-driven decisions, especially in public services, are transparent and interpretable. Individuals affected by an AI decision should have the right to an explanation.

Bridge the Skill Gap and Digital Divide: Initiatives like Digital India Mission, INDIAai FutureSkills and Responsible AI for Youth must be expanded to provide AI education, especially in Tier 2 and Tier 3 cities and rural areas.

• The World Economic Forum predicts that AI could generate 40 million new jobs in India by 2030. To grab this opportunity, focus on providing adequate training opportunities to equip workers with the skills necessary for emerging jobs.

Strengthen Cybersecurity: Develop AI-enabled cybersecurity strategies to counter AI-powered threats and protect its critical digital infrastructure.

What India can learn from other countries

The European Union (EU): The EU AI Act is the world's first comprehensive regulatory framework for AI and offers several learning points.

• Risk-Based Approach: It categorizes AI systems into unacceptable risk (banned), high-risk (strict requirements), limited risk (transparency obligations), and minimal risk (no obligations).
• India can adopt a similar risk-based model to avoid over-regulating benign applications while strictly governing critical ones.

The United States: Adopted a more sector-specific approach. The SR-11-7 regulation, for example, provides a strong governance standard for financial models in banking. This sectoral model could be useful for regulating AI in specific high-stakes domains in India.

Canada: Directive on Automated Decision-Making outlines how government departments should use AI. It uses a scoring system to assess the level of human intervention and peer review required for an AI tool, offering a practical model for ensuring oversight in public sector AI deployment.

Singapore: Framework for generative AI and uses AI effectively for urban planning and public transport, showcasing how to apply AI for public good within a governed structure.

Conclusion

India must balance innovation with responsibility. Technology should serve people, not replace human judgment completely. Success requires cooperation between government, citizens, and technology companies.

Source: INDIAN EXPRESS