India Pushes Copyright Reform at AI Summit to Enable Accessible Content
Updated (3 articles)
AI Impact Summit Highlights India’s Copyright Challenges The India‑AI Impact Summit 2026 opened in New Delhi on 19 February 2026, drawing attention to how India’s existing copyright framework restricts access, creativity and technological progress, and prompting calls for legislative reform [1]. Organisers positioned the summit as a platform for policymakers, technologists and disability advocates to discuss needed changes [1]. The event’s focus reflects growing pressure on the government to modernise outdated rules [1].
Visually Impaired Reader Blocked by Existing Law Nirmita, a visually impaired colleague, could not legally purchase a U.S. book in the DAISY (Digital Accessible Information System) format because Indian law does not permit cross‑border distribution of accessible‑format copies without publisher consent [1]. Her case illustrates how rigid copyright provisions deny disability‑friendly access to essential reading material [1]. Advocates cite her experience to argue for broader exceptions for accessible formats [1].
Marrakesh Treaty Provides Model for Accessible‑Format Exceptions Decades of lobbying by NGOs and international disability coalitions secured the Marrakesh Treaty, which allows the cross‑border exchange of accessible‑format books and mandates national exceptions when publishers fail to provide them [1]. The treaty serves as a benchmark for India’s reform proposals, highlighting the benefits of legally sanctioned accessibility measures [1]. Proponents argue that adopting similar provisions would align India with global best practices [1].
Current Copyright Act Gives Near‑Perpetual Monopoly India’s Copyright Act of 1957 automatically vests copyright the moment a work is created and protects it for the author’s life plus 70 years, effectively granting a near‑perpetual monopoly over digital creations [1]. This long term protection turns everyday digital content into protected works for centuries, discouraging reuse and innovation [1]. Critics contend that the statute has not kept pace with the rapid evolution of AI and digital media [1].
Regional Study Shows AI Training Mostly Illegal A LIRNEasia study of seven South‑ and Southeast‑Asian nations found that, aside from the Philippines, Sri Lanka and India’s narrow “transient storage” exception, copyright law makes web crawling and AI model training unlawful [1]. The research underscores the legal uncertainty faced by AI developers across the region [1]. India’s limited exception is deemed insufficient for large‑scale AI training activities [1].
Absence of Broad Text‑and‑Data‑Mining Exception Creates Uncertainty Without a comprehensive text‑and‑data‑mining (TDM) exception, Indian AI developers confront significant legal risk when training models on copyrighted material [1]. The article’s author urges the introduction of safe‑harbour provisions and support for open‑licensed datasets to foster innovation [1]. Such reforms are presented as essential for India to remain competitive in the global AI landscape [1].
Timeline
1957 – India’s Copyright Act automatically vests copyright the moment a work is created and extends protection for the author’s life plus 70 years, turning everyday digital creations into near‑perpetual monopolies and limiting reuse for AI training [1].
2010s – The Marrakesh Treaty, secured after years of disability‑rights lobbying, permits cross‑border exchange of accessible‑format books and allows national exceptions for converting works when publishers fail to provide them, setting a precedent for accessibility‑focused copyright flexibilities [1].
Dec 10‑11, 2025 – The Department for Promotion of Industry and Internal Trade (DPIIT) publishes a working paper proposing a mandatory licensing framework for data‑mining, creating a non‑profit body to collect royalties from AI firms that scrape public Indian content and linking payments to the revenue benefits those firms derive [3].
Dec 23, 2025 – New York Times reporter John Carreyrou and five co‑authors file a federal lawsuit in California against Google, xAI, OpenAI, Anthropic, Meta and Perplexity, alleging the companies used their copyrighted books without permission to train large‑language‑model chatbots [2].
Dec 23, 2025 – The complaint becomes the first copyright case to name Elon Musk’s xAI as a defendant, highlighting the expanding legal exposure of emerging AI firms over training‑data practices [2].
Dec 23, 2025 – Plaintiffs reject a class‑action strategy, arguing that statutory caps would yield only a tiny fraction of potential damages and citing Anthropic’s earlier $1.5 billion settlement as evidence that per‑work payouts are insufficient [2].
Feb 18, 2026 – The India‑AI Impact Summit 2026 convenes in New Delhi, drawing attention to India’s copyright dilemma; organizers cite the case of Nirmita, a visually impaired reader blocked from legally purchasing a DAISY‑format book, as a vivid illustration of how current law hampers disability‑friendly access [1].
Feb 2026 – A LIRNEasia study presented at the summit finds that AI training is illegal in most South‑ and Southeast‑Asian nations except for limited “transient storage” exceptions in the Philippines, Sri Lanka and India, underscoring regional legal uncertainty for AI developers [1].
Feb 2026 – Summit participants call for safe‑harbour provisions and the creation of open‑licensed datasets to address India’s lack of a text‑and‑data‑mining exception, urging copyright reform that balances creator remuneration with AI innovation [1].