AI Bias, International Law, and India’s Need for a Sovereign AI Stack
Introduction
Artificial Intelligence (AI) is rapidly becoming a key tool used by policymakers, analysts, and researchers to interpret global events. However, AI systems are not neutral. They often reflect the biases of the data on which they are trained, which currently comes largely from Western sources.
This raises a crucial question for India: Should it rely on foreign AI systems or build its own sovereign AI architecture?
For UPSC aspirants, this topic is relevant to GS-II (International Relations), GS-III (Science & Technology), and Ethics/Essay topics on technology and geopolitics.
The Legal Debate: Military Activities in the EEZ
The dispute revolves around the interpretation of the United Nations Convention on the Law of the Sea (UNCLOS).
Under UNCLOS, countries have special rights in their Exclusive Economic Zone (EEZ), which extends 200 nautical miles from their coast.
Article 58 of UNCLOS
Article 58 of UNCLOS grants foreign states freedoms such as:
-
navigation
-
overflight
-
other lawful uses of the sea related to these freedoms
However, the interpretation of this article is deeply contested.
Two Competing Interpretations
1. Western Interpretation
Countries like the United States interpret Article 58 broadly.
They believe that foreign states can conduct military activities in another country’s EEZ, including:
-
naval exercises
-
submarine operations
-
intelligence gathering
-
weapons testing
As long as these occur outside territorial waters, they are considered legal.
2. Global South Interpretation
Countries such as:
-
India
-
China
-
Indonesia
-
Brazil
-
South Africa
argue that military activities in an EEZ require the consent of the coastal state.
They rely on Article 58(3), which requires states to show “due regard” for the rights of coastal states.
Thus, the legality of military operations in EEZs remains unresolved in international law.
The Humanitarian Law Dimension
Another important aspect relates to humanitarian obligations under the Second Geneva Convention.
Article 18
Article 18 of the Second Geneva Convention requires parties to:
-
search for survivors
-
rescue shipwrecked sailors
-
provide assistance without delay
Reports indicate that after the sinking of the Iranian warship, the attacking submarine left quickly, and rescue operations were conducted by Sri Lanka.
This raises further legal and humanitarian questions that the AI system initially failed to consider.
AI Bias: A Structural Problem
The incident demonstrates that AI systems often reflect structural bias.
Why bias occurs
Most AI models are trained on:
-
Western academic publications
-
Western legal scholarship
-
Western media sources
As a result:
-
Western interpretations appear as default truths
-
Global South perspectives become secondary or invisible
Thus, AI outputs may reinforce global power asymmetries.
The Global AI Divide
The global AI ecosystem is increasingly dominated by two major powers:
1. The U.S. AI ecosystem
Led by companies such as OpenAI, Google, and Microsoft.
Strengths:
-
advanced chips
-
powerful cloud infrastructure
-
cutting-edge AI models
2. The Chinese AI ecosystem
Led by firms such as Baidu, Alibaba, and Tencent.
China’s AI stack is closely integrated with state policy and data governance.
India’s Strategic Choice
India now faces three possible options.
Option 1: Adopt the U.S. AI stack
Advantages:
-
faster access to technology
-
global research collaboration
Risks:
-
dependence on foreign infrastructure
-
Western biases embedded in AI systems
Option 2: Adopt the Chinese AI ecosystem
This option raises concerns about:
-
security
-
geopolitical alignment
-
data sovereignty
Option 3: Build a Sovereign Indian AI Stack
This approach involves developing:
-
domestic computing infrastructure
-
indigenous datasets
-
multilingual AI models
-
national data governance frameworks
India has already shown its ability to build large-scale digital infrastructure through initiatives like:
-
Aadhaar
-
Unified Payments Interface
A similar strategy could be applied to AI.
Risks of AI Dependence
If India relies entirely on foreign AI models, it could lead to digital colonialism, where:
-
algorithms control knowledge production
-
foreign platforms shape public discourse
-
national perspectives are underrepresented
This would affect not only technology but also geopolitics and strategic thinking.
The Way Forward for India
India must adopt a balanced strategy:
-
Invest in domestic AI research and computing infrastructure
-
Develop Indian-language datasets and models
-
Build secure national data ecosystems
-
Collaborate internationally while maintaining technological sovereignty
The goal is integration without dependence.
Why This Matters for UPSC
This issue intersects several parts of the UPSC syllabus.
GS-II
-
International law
-
India’s foreign policy
-
Global governance
GS-III
-
Artificial Intelligence
-
Technology and national security
-
Data governance
Essay Topics
Possible themes include:
-
Technology and geopolitical power
-
Data sovereignty and digital colonialism
-
Ethics of artificial intelligence
Conclusion
AI is no longer just a technological tool — it is becoming an architecture of interpretation, shaping how events, laws, and geopolitical realities are understood.
Countries that rely entirely on foreign AI systems may eventually think through someone else’s frameworks.
For India, the challenge is clear: it must transition from being a consumer of AI technologies to a creator of AI knowledge systems.
Just as India developed its space programme, nuclear capabilities, and digital public infrastructure, it must now build a sovereign AI stack that reflects its own linguistic, cultural, and strategic realities.
No comments:
Post a Comment