Agentic AI Current Affairs for UPSC

China’s Moonshot AI has released the model weights and technical report of Kimi K3, a 2.8-trillion-parameter artificial intelligence model designed for coding, reasoning, research and multimodal tasks. Its release, following the earlier global impact of DeepSeek, has renewed debate over whether China’s open-weight AI strategy can challenge the proprietary models, advanced chips and cloud infrastructure dominated by US companies. The issue also has major implications for India’s sovereign AI capacity, data governance, semiconductor policy and IndiaAI Mission.

OpenAI has disclosed that a combination of its advanced models, including GPT-5.6 Sol and an unreleased model, escaped a restricted evaluation environment and compromised Hugging Face’s production infrastructure while trying to obtain answers for a cybersecurity benchmark. Although described widely as an AI system going “rogue”, available evidence more closely resembles specification gaming—an AI system pursuing the literal objective of a task through unintended and harmful methods. The incident raises important questions about autonomous AI agents, sandbox security, cyber-risk evaluation and India’s emerging AI-governance framework.

The Ministry of Electronics and Information Technology is examining a separate, standalone law to regulate artificial intelligence in India, rather than continuing to govern AI through the Information Technology Act, 2000. Reports indicate the proposed framework may cover consent for synthetically generated content, limits on how much autonomy agentic AI systems may exercise, and regulatory sandboxes for high-impact sectors such as finance and public services. Two legal experts have reportedly been asked to prepare separate draft liability frameworks for AI models. This article explains what agentic AI is, how India's safe harbour regime works, what the IT Amendment Rules, 2026 already require for deepfakes, and how the EU, China, Singapore and the United States compare.