"AI Governance Debate: Human Agency, Safety, and Geopolitical Control"
Large Language Model (LLM)
AI system that predicts text by assigning probabilities to word sequences using mathematical optimisation.
Human in the loop
principle requiring a human to oversee and be able to override AI decisions at key points.
Artificial intelligence, at its core, aims to simulate human cognitive functions in machines. Large Language Models (LLMs), the technology behind today's most capable AI systems, work by assigning probabilities to sequences of words and using mathematical optimisation to minimise prediction errors during training. This makes them powerful pattern-recognisers. It also means they can produce biased or incorrect outputs, since they learn from human-generated data that itself contains errors and prejudices. Understanding this limitation is essential to any serious governance discussion.
The international community has attempted to build shared governance frameworks through a series of summits. The Bletchley Park AI Safety Summit (2023) was the first major multilateral effort, bringing together governments and frontier AI labs to discuss existential and near-term risks. It was followed by the Seoul AI Summit (2024). The AI Action Summit held in Paris in early 2025 was co-chaired by India and France, building on these earlier milestones. Notably, the United States and the United Kingdom did not sign the Paris declaration on inclusive and sustainable AI, reflecting a broader divergence: the US has favoured innovation and private-sector leadership, while Europe has emphasised risk-based regulation and fundamental rights. China has embedded AI within national strategy and state sovereignty. These three philosophies now compete as much as the technologies themselves.
India's own position has evolved. The IndiaAI Mission and the National Strategy for Artificial Intelligence both emphasise responsible, sovereign, and inclusive AI development. India's co-chairmanship of the Paris summit signals its ambition to be a rule-shaper, not merely a rule-taker, in global AI governance.
The central governance challenge is not simply whether AI systems are dangerous in some abstract sense. It is whether humans retain the practical ability to understand, audit, and stop AI systems when they go wrong. This is captured in the concept of meaningful human oversight, sometimes called the "human in the loop" principle. The problem is that this phrase can become a legal fiction. A person who cannot understand a system's reasoning, cannot test its conclusions, and cannot intervene quickly enough may remain formally responsible while exercising no real control.
As AI systems move from generating text to taking autonomous actions, in cybersecurity operations, financial decisions, infrastructure management, the stakes of this gap rise sharply. The concern is not that AI will become conscious or malevolent. It is that a system pursuing an objective with greater autonomy and persistence than its designers fully understand may cause harm through capability, access, and poorly constrained objectives alone. …
Consider the following statements regarding the AI Action Summit held in Grand Palais, Paris in February 2025:
I.Co-chaired with India, the event builds on the advances made at the Bletchley Park Summit held in 2023 and the Seoul Summit held in 2024.
II.Along with other countries, US and UK also signed the declaration on inclusive and sustainable AI.
Which of the statements given above is/are correct?
Which of the following statements with regard to Large Language Models (LLMs) used in machine learning is/are correct?
1.LLMs assign probabilities to the next possible word and then pick the one with the highest probability.
2.LLMs process data through mathematical optimization to minimise prediction errors.
3.LLMs produce unbiased outputs.
Select the answer using the code given below:
With the present state of development, Artificial Intelligence can effectively do which of the following?
1.Bring down electricity consumption in industrial units
2.Create meaningful short stories and songs
3.Disease diagnosis
4.Text-to-Speech Conversion
5.Wireless transmission of electrical energy
Select the correct answer using the code given below:
Consider the following statements regarding international AI governance summits:
1.The Bletchley Park AI Safety Summit was held in 2023 and was the first major multilateral summit on AI safety risks.
2.The AI Action Summit held in Paris in 2025 was co-chaired by India and France.
3.The United States and the United Kingdom signed the Paris declaration on inclusive and sustainable AI.
4.The Seoul AI Summit was held between the Bletchley Park and Paris summits.
How many of the above statements are correct?
Which of the following statements with regard to Large Language Models (LLMs) is/are correct?
1.LLMs assign probabilities to sequences of words as part of their text generation process.
2.LLMs use mathematical optimisation to minimise prediction errors during training.
3.LLMs produce outputs that are free from bias because they are trained on large datasets.
Select the correct answer using the code given below:
Match the following AI governance frameworks with their primary regulatory philosophy:
List I (Country/Bloc) — List II (Primary AI Regulatory Philosophy)
A.United States — 1. State-directed AI embedded in national strategy and information sovereignty
B.European Union — 2. Innovation-led, private-capital-driven, technological leadership
C.China — 3. Risk-based regulation emphasising consumer protection and fundamental rights
Select the correct match:
With the present state of development, Artificial Intelligence can effectively perform which of the following tasks?
1.Disease diagnosis support
2.Text-to-speech conversion
3.Wireless transmission of electrical energy
4.Optimising electricity consumption in industrial units
Select the correct answer using the code given below:
Consider the following statements about AI governance and the concept of 'meaningful human oversight':
1.A person who cannot understand an AI system's reasoning but remains legally responsible for its decisions exercises meaningful human oversight.
2.Effective AI governance requires that a system's behaviour can be independently audited and that access to it can be withdrawn.
3.Dependence on foreign AI infrastructure can erode a nation's practical sovereignty even when formal sovereignty is legally intact.
Which of the statements given above is/are correct?
Want the full note, Prelims Facts & MCQ answers?
Create a free account — takes less than a minute.
Create free account