HUMAN-CENTRIC ARTIFICIAL INTELLIGENCE: BUILDING AN ETHICAL, INCLUSIVE AND RESPONSIBLE AI ECOSYSTEM

Why In News?

Global leaders demand an enforceable global Human-Centric regulatory regime to mitigate AI-induced job volatility, data privacy threats, and the proliferation of misinformation.

What is Human-Centric AI?

Human-Centric AI is an approach that prioritizes human needs, ethical values, and societal well-being in the design and deployment of AI systems. 

Rather than focusing solely on raw automation and machine logic, it emphasizes human-in-the-loop collaboration, ensuring AI enhances human capabilities while protecting individual rights.

Core Principles

Augmentation over Replacement: AI should assist, not replace, human judgment and creativity (e.g., a doctor using AI to analyze scans rather than replacing the doctor).  

Ethics and Fairness: Systems must be designed to mitigate algorithmic bias and discrimination. 

Transparency & Explainability: The "black box" nature of AI should be minimized so users can understand how and why a decision was made. 

Meaningful Human Control: There must always be accountability and human oversight in critical decision-making (e.g., healthcare, criminal justice). 

AI as a Tool to Augment Human Capabilities: Artificial Intelligence acts as a "force multiplier" rather than a replacement for human intellect.

• Governance and Administration: AI assists in real-time citizen grievance redressal, predictive policing, and optimizing public policy by analyzing massive datasets. 
• Healthcare: Tools assist doctors with medical imaging and diagnostics, enabling faster and more accurate treatments while preserving the human touch in patient care.
• Economy and Industry: AI boosts productivity across sectors like agriculture (pest prediction, precision farming) and manufacturing (predictive maintenance).
• Education: AI drives personalized learning. AI acts as a research assistant, organizing current affairs or providing rapid feedback.

Opportunities Offered by Artificial Intelligence

Healthcare and Precision Medicine: Machine learning analyzes X-rays and MRIs to detect anomalies early, while the Indian AI healthcare market targets $35 billion by 2030.

Education and Personalized Learning: AI creates dynamic K-12 learning pathways and supports special education through platforms like YUVAi.

Agriculture and Smart Farming: AI drones monitor crop health, and apps like Kisan e-Mitra optimize sowing and irrigation schedules based on real-time data.

Governance and Public Service Delivery: The e-Courts Project Phase III uses AI for translating Supreme Court judgments, while SabhaSaar automates Gram Panchayat meeting minutes.

Scientific Research and Innovation: Innovations like AlphaFold solve complex protein-folding problems, paving the way for medical breakthroughs.

Climate Change Mitigation: The India Meteorological Department (IMD) uses AI models like the Advanced Dvorak Technique and MausamGPT to forecast extreme weather events.

Industrial Productivity and Economic Growth: Generative AI holds the potential to increase global GDP by 7% by automating repetitive workflows.

Financial Inclusion: AI algorithms instantly analyze vast financial datasets to detect fraudulent transactions and offer robo-advisory services.

Risks Associated with AI

Algorithmic Bias and Discrimination: AI models reproduce historical caste, gender, and racial prejudices, often penalizing marginalized groups during hiring or loan approvals.

• A large-scale UN Women study analyzing 133 AI systems revealed that 44% of systems demonstrated gender bias. Large Language Models (LLMs) associate women with domestic roles and men with career or leadership roles.

Job Displacement and Labour Market Disruptions: Goldman Sachs economists project that AI could displace roughly 15 million US workers (about 9% of the workforce) over a 10-year transition.

• The World Economic Forum estimates that by 2030, AI and information processing technologies will displace 92 million jobs, but will simultaneously create 170 million new roles, resulting in a net positive workforce change.  

Privacy and Data Protection Concerns: Training Large Language Models (LLMs) requires extracting massive amounts of Personally Identifiable Information (PII) without transparent consent.

Deepfakes and Misinformation: Malicious actors use deepfakes to override facial recognition security and spread targeted political misinformation.

• An investigation by Forbes into a $138.5 million financial fraud case highlighted how AI face-swapping technology successfully deceived real-time liveness detection systems.

Cybersecurity Threats: AI-enabled phishing campaigns and automated hacking tools severely exacerbate national geopolitical risks.

• CrowdStrike 2026 Global Threat Report documents an 89% increase in attacks by AI-enabled adversaries, highlighting the exploitation of legitimate AI tools to generate malicious commands.  

Autonomous Weapons: The military deployment of super-intelligent, unaccountable lethal weapons poses grave global security risks.

• Intelligence Warning: The Five Eyes Intelligence Alliance issued a joint warning detailing how frontier AI models capable of overwhelming government defenses could emerge within months rather than years.  

Concentration of Technological Power: The costs of AI development—such as the $191 million spent training Gemini Ultra—consolidate power within a few massive tech corporations.

Environmental and Resource Strain: The sprawling data centers required for AI model training consume land, water, and energy, threatening global climate goals. 

• According to the International Energy Agency (IEA) Energy and AI Report, global data center electricity consumption could double from 415 TWh to roughly 945 TWh by 2030, which would account for around 3% of total global electricity.  
• The United Nations University reports that AI data centers could consume 9.3 trillion liters of water annually by 2030—roughly the annual domestic water needs of 1.3 billion people.  

Ethical Principles for Responsible AI

Transparency: Users must explicitly know when they interact with AI or consume synthetically generated deepfake content.

Accountability: Regulations must clearly identify the developer, deployer, or user legally responsible when an algorithm denies a citizen their rights.

Explainability: AI systems must provide understandable reasoning for their decisions, especially in credit scoring and legal judgments.

Fairness: Systems must actively filter out discriminative proxies related to caste, gender, and regional backgrounds.

Privacy by Design: Developers must strictly limit data collection, complying with frameworks like India's Digital Personal Data Protection (DPDP) Act.

Safety and Reliability: High-stakes systems (e.g., autonomous driving or predictive policing) require fail-proof reliability frameworks because errors are lethal.

Human Oversight: The EU AI Act strictly prohibits fully automated decision-making in high-risk categories, ensuring a "human-in-the-loop" safeguard,.

Contextual Integrity: AI must respect local cultural norms and languages, rather than imposing generic Western fairness models on complex societies like India.

Global AI Governance Framework

United Nations Initiatives: The UN drives global discussions to prevent fragmentation and align AI development with Sustainable Development Goals.

UNESCO Recommendation on the Ethics of AI: The 2021 UNESCO framework establishes a global baseline for integrating ethical reflection directly into AI lifecycles.

OECD AI Principles:  The OECD provides the globally accepted definition of AI systems, establishing uniform legal interpretations across borders.

G7 Hiroshima AI Process: G7 nations collaborate to craft interoperable AI standards that mitigate systemic risks and promote safe cross-border data flows.

International Scientific Panel on AI: Proposed global bodies (similar to the IPCC for climate) to independently audit advanced foundational models.

The European Union AI Act: The EU AI Act divides AI into unacceptable, high, limited, and minimal risk tiers, banning behavioral manipulation and mandating rigorous audits,,.

Global Partnership on AI (GPAI): An international OECD-hosted initiative, founded with India, supporting responsible and human-centric AI development

India's Approach to Human-Centric AI

IndiaAI Mission: The Union Cabinet approved a Rs 10,372 crore IndiaAI Mission to establish indigenous compute capacity, datasets, and innovation ecosystems.

AI for All Vision: Launched by NITI Aayog in 2018, this strategy deploys AI to bridge gaps in healthcare, agriculture, and education.

Digital Public Infrastructure (DPI): Platforms like the Digital ShramSetu Mission assist over 490 million informal workers by scaling technological access.

National Data Governance Framework: Platforms like AIKosh act as repositories housing over 7,500 open-source datasets to lower entry barriers for developers.

Responsible AI Guidelines: The 2025 MeitY guidelines outline seven guiding "Sutras" to ensure AI is transparent, fair, and accountable in rural service delivery.

Sovereign AI Models: Development of sovereign LLMs like BharatGen (supporting 22 languages) and startups like Sarvam AI challenge global monopolies with culturally accurate models.

Key Challenges

Regulatory Gaps: The DPDP Act 2023 focuses strictly on data privacy but lacks mechanisms to audit or punish algorithmic discrimination.

Digital Divide: Half of India lacks robust internet access, entirely erasing them from the datasets that train national AI models.

Lack of AI Literacy: Transforming a legacy workforce to adopt AI requires massive capital and training, which many small businesses lack.

Cross-Border Governance: The borderless nature of generative AI makes it difficult to enforce national laws uniformly on international tech conglomerates.

Data Sovereignty: Heavy reliance on Western foundational models severely threatens India's strategic digital autonomy and cultural preservation.

Ethical Enforcement: Current corporate ethical frameworks are voluntary; without stringent legal penalties, companies prioritize profit over safety.

Environmental and Compute Costs: India severely lacks localized semiconductor manufacturing and data-center density, hindering native large-scale model training.

Way Forward

Establishing a Human-Centric AI Governance Framework: India must enact risk-tiered, legally enforceable AI laws equipped with strict auditing mechanisms and clear liability chains.

Strengthening Global Cooperation: India must use platforms like GPAI and the UN to shape interoperable global rules that protect the specific interests of the Global South.

Investing in AI Ethics Research: The state must fund dedicated research exploring how AI intersects with Indian social hierarchies, actively designing bias-mitigation tools.

Building AI Skills and Digital Literacy: Rapidly scale initiatives like FutureSkills PRIME to retrain the existing labour force for augmented human-machine collaboration.

Ensuring Algorithmic Transparency: Public institutions must mandate independent lifecycle audits and "human-in-the-loop" safeguards for any AI affecting fundamental rights.

Promoting Inclusive Innovation: Sovereign Solutions: Accelerate funding for localized, multilingual models like BharatGen to ensure AI tools inherently understand India's vernacular and cultural realities.

Encouraging Multi-Stakeholder Participation: AI governance must integrate civil society, legal academics, and marginalized communities to validate real-world algorithm impacts.

Disaggregated Data Collection: Public bodies must gather representative datasets mapped across caste, gender, and region to prevent algorithmic exclusion.

Conclusion

To realize AI's vast potential, nations must reject voluntary self-policing and urgently implement enforceable, human-centric legislation that subordinates technological power to constitutional dignity, privacy, and equity.

Source: THEHINDU