What Happened
- The European Commission is assessing whether to designate OpenAI's ChatGPT as a "Very Large Online Search Engine" (VLOSE) under the EU's Digital Services Act (DSA)
- OpenAI disclosed that ChatGPT's search feature averaged 120.4 million monthly active users in the EU — far exceeding the DSA's 45 million-user threshold for special designation
- A Commission spokesperson confirmed regulators are "currently assessing the information" to determine if ChatGPT meets the criteria for designation
- Designation would subject ChatGPT to stricter regulatory requirements including enhanced transparency, algorithmic accountability, and risk management obligations
- The move represents the most significant potential expansion of AI platform regulation under European law to date
Static Topic Bridges
The EU Digital Services Act (DSA): Framework and Scope
The Digital Services Act (DSA), which entered into force in November 2022 and became fully applicable in February 2024, is the EU's comprehensive framework for regulating digital intermediaries — platforms, online marketplaces, social networks, and search engines. It creates a tiered regulatory system based on user scale, with the heaviest obligations falling on "Very Large Online Platforms" (VLOPs) and "Very Large Online Search Engines" (VLOSEs).
- DSA threshold for VLOP/VLOSE designation: 45 million monthly active users in the EU
- VLOP/VLOSE obligations: Detailed risk assessments for systemic risks (illegal content, fundamental rights), independent audits, transparency reports, data access for researchers, interoperability requirements, and crisis response protocols
- Enforcement: European Commission directly oversees VLOPs/VLOSEs; national Digital Services Coordinators handle smaller platforms
- Penalties for non-compliance: Up to 6% of global annual turnover; repeat violations can lead to temporary suspension in the EU
- DSA complements the EU AI Act (passed 2024), which separately regulates high-risk AI applications
Connection to this news: ChatGPT's 120.4 million monthly EU users far exceed the 45 million threshold, making it a strong candidate for VLOP/VLOSE designation — which would impose extensive compliance obligations on OpenAI for the first time under European platform law.
AI Governance and the EU AI Act
The EU has pursued the world's most comprehensive AI regulatory framework. The EU AI Act (adopted June 2024) classifies AI systems by risk level — prohibited, high-risk, limited-risk, and minimal-risk — and imposes proportionate obligations. Large language model providers like OpenAI face transparency and systemic-risk obligations under the Act's General-Purpose AI (GPAI) provisions.
- EU AI Act: Risk-based framework — banned uses (biometric mass surveillance, social scoring), high-risk (healthcare, education, employment, critical infrastructure), and general-purpose AI (GPAIs like GPT-4)
- GPAI obligations: Technical documentation, copyright compliance, transparency about training data; for systemic-risk GPAIs (those trained on >10^25 FLOPs), also adversarial testing and incident reporting
- DSA and AI Act are complementary but distinct: DSA governs platform conduct (how the service operates), AI Act governs AI model capabilities and deployment risks
- "Simple bias tests" will not be sufficient for compliance — DSA requires comprehensive risk management and ongoing audit
Connection to this news: Designating ChatGPT under both the DSA and AI Act would create a dual compliance burden for OpenAI — platform-level obligations (DSA) and model-level obligations (AI Act) — making the EU the world's most regulated environment for AI deployment.
India's Approach to AI Regulation: Contrast with the EU
While the EU has opted for comprehensive, binding AI regulation, India has pursued a softer, principles-based approach. India's National AI Strategy (NITI Aayog, 2018) and subsequent AI governance frameworks have prioritised innovation and adoption over pre-emptive regulation. The Digital India Act (DIA), expected to replace the legacy IT Act 2000, is expected to address some platform accountability issues but falls short of the EU's detailed prescriptive regime.
- India's IT Act 2000 (Section 69A, 79): Provides intermediary liability exemptions and takedown mechanisms but does not impose proactive algorithmic risk assessment obligations
- NITI Aayog's "Responsible AI" framework (2021): Voluntary principles (fairness, accountability, transparency, privacy) without binding enforcement
- Digital Personal Data Protection Act (DPDPA, 2023): India's first data protection law — applies to personal data processing, complementary to but distinct from AI-specific regulation
- India's stance at global AI forums: Advocates for AI regulation that enables developing country access to technology rather than creating regulatory barriers that entrench incumbents
- G20 New Delhi Declaration (2023): Endorsed voluntary AI principles and called for interoperable frameworks — reflecting India's preference for soft governance
Connection to this news: The EU's move to regulate ChatGPT under the DSA represents the opposite regulatory philosophy from India — prescriptive vs. principles-based, mandatory vs. voluntary. For UPSC, this contrast is highly relevant to debates on digital governance and India's digital policy choices.
Key Facts & Data
- ChatGPT monthly active users in EU: 120.4 million (on average over past 6 months) — vs. 45 million DSA threshold
- DSA VLOP/VLOSE penalty: up to 6% of global annual turnover for non-compliance
- DSA: Entered into force November 2022; fully applicable February 2024
- EU AI Act: Adopted June 2024; GPAI systems trained on >10^25 FLOPs face systemic-risk obligations
- India's DPDPA: Passed 2023 (data protection); Digital India Act expected to replace IT Act 2000
- OpenAI's ChatGPT: Released November 2022; reportedly crossed 300 million weekly active users globally by early 2025
- European Commission: Direct regulator of designated VLOPs/VLOSEs under DSA