Panel Discussion: AI and the Creative Economy | India AI Impact Summit

Contents

Executive Summary

This panel explores the intersection of artificial intelligence and creative industries, examining whether AI strengthens or weakens cultural diversity, the adequacy of global IP frameworks for AI-generated content, and practical solutions for balancing creator rights with technological innovation. Key consensus emerges around human-centered governance, the need for consent and attribution mechanisms, and India's strategic opportunity to leverage its vast public domain cultural heritage (the Ethas—epic traditions like the Mahabharata and Ramayana) in AI training datasets.

Key Takeaways

India's window of opportunity is now: India can digitize its public domain cultural heritage using AI tools (OCR, voice models via Sarvam AI) and incorporate it into AI training datasets, positioning itself strategically before Western AI models achieve dominance. This requires creative work first, not AI work.
Technological solutions over legal treaties: Global IP harmonization is unrealistic in the short term (194 WIPO members require consensus). The practical path forward is technical infrastructure—metadata standards, opt-in/opt-out mechanisms, and transparency tools that allow both creators and tech companies to navigate AI responsibly.
Humans must remain central: Storytelling, creativity, and human collaboration are irreplaceable. The future is not about AI replacing creators but about AI as a tool that enhances human creativity—provided consent, attribution, and compensation mechanisms are built in.
Creator agency is collapsing without intervention: Without consent mechanisms and agency over how their work is used, creators are retreating from open sharing and collaborative commons, which paradoxically harms everyone by shrinking the knowledge base AI can learn from.
Middle-ground licensing is essential: Between "no" (restrictive copyright) and "yes" (unrestricted use), there must be nuanced conditional permissions: "yes if you attribute," "yes if you compensate," "yes if you support open infrastructure," negotiated through evolving Creative Commons-like frameworks.

Key Topics Covered

AI's impact on creative industries: Production music (42–48% AI-generated in broadcast), scriptwriting, character design, visual effects, and gaming
Cultural diversity and representation: Risk of weakening unless governed by open-source principles and transparency
Intellectual property frameworks: Current global copyright systems and their adequacy for AI-generated content
Consent and attribution: Ethical inconsistencies between artists using AI freely while objecting to their work being used in AI training
India's strategic advantage: Digitization and AI representation of public domain cultural heritage
Technological vs. legal solutions: WIPO's pragmatic approach focusing on technical infrastructure rather than international treaties
Creator compensation: Balancing remuneration for creators with tech industry utilization of AI models
Shrinking commons problem: Risk of creators withdrawing work from public sharing due to lack of consent mechanisms
Human creativity as irreplaceable: The enduring role of storytelling and human creativity in the AI era

Key Points & Insights

Data representation gap: India's rich oral traditions and epic heritage are severely underrepresented in AI training datasets compared to Hollywood or AAA gaming content, despite representing 20% of global population. This creates both a risk and an opportunity.
Public domain leverage: The Ethas (Mahabharata, Ramayana, Bhagavad Gita) exist in the public domain as living traditions, offering a unique asset for AI training without copyright complications—no one has created significant IP atop them yet at scale.
Speed of AI vs. human learning: The fundamental difference between how humans learn through imitation (which takes time) versus how AI generates outputs at machine speed necessitates new legal and ethical boundaries.
Consent as core ethical issue: The central ethical inconsistency is that some creators use AI trained on unconsented global creative works while simultaneously objecting to their own work being scraped. This highlights the gap between technological capability and ethical governance.
Shrinking commons paradox: Creators are withdrawing works from public sharing and CC licenses due to lack of agency over AI usage, which undermines the collaborative human knowledge-building process that has historically relied on copyright clarity and open licensing.
WIPO's pragmatic approach: Rather than pursuing international legal harmonization (requiring consensus among 194 member states), WIPO is exploring technological infrastructure solutions—opt-in/opt-out mechanisms, metadata for attribution, and technical ways to signal creator intent.
Scale changes everything: AI's massive scale transforms margin problems (5% edge cases in open knowledge communities) into central problems (95% of creators now affected), requiring fundamentally rethinking governance approaches.
Norms over regulation: President Macron's framing of AI governance as a "civilization question" rather than purely regulatory—what kind of society do we want regarding creator recognition and data rights?
Collaboration, not replacement: Despite automation concerns, creative production (games, films, stories) remains fundamentally collaborative, involving hundreds of people, and is unlikely to be fully automated by AI agents in the near term.
Attribution and credit mechanisms needed: There's broad consensus that while knowledge builds on prior knowledge, the technological layer of AI-generated fusion of styles/inspirations requires some form of credit attribution—a middle ground between open use and full restriction.

Notable Quotes or Statements

Speaker	Quote	Context
Nicholas (Tara Gaming)	"India with its phenomenal history and culture and 20% of the world population has to punch at its weight in the training data sets of these AI models—and the first work that needs to be done is a creative work, not an AI work."	Emphasizing India's strategic leverage of cultural heritage
Nicholas (Tara Gaming)	"The bigger question is: do we want AI as a society to have the best data? The answer is probably yes—on the condition that you are open."	Conditional support for AI data access
President Macron (referenced)	"It's not about regulation, it's about civilization."	Framing AI governance as existential societal choice
Anna Tumadir (Creative Commons)	"If nobody shares then there's nothing left for us."	Warning about shrinking commons from creator withdrawal
Anna Tumadir (Creative Commons)	"There is something with the technological layer where if you are fusing together different things or asking for certain styles or certain inspirations, there really should be some sort of form of credit given."	Advocating middle-ground attribution requirements
Ken (WIPO)	"The views are very much different... It's not something like zero and one and let's see 0.5. It's not like mathematics."	Acknowledging fundamental disagreement among stakeholders
Ken (WIPO)	"Our idea is let's think about something more practical, more technological... that creators can be benefited or remunerated as well as tech industries can utilize those products."	WIPO's pragmatic pivot from legal to technical solutions
Nicholas (Tara Gaming)	"The creative process is above language, above image... and everybody agrees in the AI community there's two three things that need to happen to reach AGI and those overlap with creativity that makes us unique."	Positioning human creativity as foundational to AGI

Speakers & Organizations Mentioned

Speaker	Title/Role	Organization
Rammit Da	Moderator	(Not specified; appears to be director of panel)
Nicholas Granitino	Chairman	Tara Gaming
Kenny Chiro Natsum	Assistant Director General	WIPO (World Intellectual Property Organization)
Anna Tumadir	Chief Executive Officer	Creative Commons
Nir Troy	Managing Director (Keynote speaker)	Hungama Digital Media Entertainment
(President) Macron	(Referenced, not present)	France
Prime Minister Modi	(Referenced, not present)	India
Holly H. Turn	Artist (mentioned as case study)	(Not specified)
Heap	Artist (mentioned as case study)	(Not specified)

Organizations Referenced:

Sarvam AI (Indian AI company with OCR and voice models)
Deep Mind (AI research)
WIPO (UN international organization for IP)
Creative Commons (open licensing organization)
European Union, China, France, India (policy contexts)

Technical Concepts & Resources

Concept/Tool	Description	Context
Latent space	The training dataset and representations within AI models; crucial for determining what outputs are generated	Nicholas emphasizes need for Indian cultural content in latent space
Foundation models	Large-scale AI models trained on massive datasets (mentioned without deep definition)	Discussed as base for generative AI in creative industries
OCR (Optical Character Recognition)	Technology for digitizing text	Sarvam AI's tool for digitizing Indian cultural texts
Opt-in/opt-out mechanisms	Technical infrastructure for creator consent signaling	WIPO's proposed solution for managing AI training permissions
Generative AI	AI systems that generate new content (music, art, text, code)	Central focus of discussion; 42–48% of broadcast production music is AI-generated or AI-assisted
CC Licenses (Creative Commons)	Open licensing framework allowing creators to specify usage rights	Discussed as model for nuanced consent; creators increasingly restricting these due to AI concerns
Protein Data Bank	Historical scientific database used in AI training (protein folding)	Nicholas uses as example of underrecognized data contributions to AI breakthroughs
Scraping/crawling/trolling	Methods of collecting training data from internet	Anna refers to how world's creativity has been aggregated without consent
Normative frameworks	Social/ethical norms rather than legal rules	Anna suggests as potential solution to copyright limitations for AI governance
Broadcast industry production music statistics	42–48% AI-generated or AI-assisted	Cited metric of AI adoption in one creative vertical

Additional Context & Structure Notes

Panel Format:

~30-minute discussion with structured opening questions, followed by conversational exploration
Three panelists covering business (Tara Gaming), policy (WIPO), and open knowledge (Creative Commons) perspectives

Key Timeline Reference:

WIPO launching first technical working meeting on March 17 (year not explicitly stated, likely 2024 or 2025)
Discussion positioned as immediate-term (next 12 months) and longer-term (decade-scale) governance challenges

Geographic Focus:

India positioned as primary case study and beneficiary of strategic positioning
Global IP frameworks (EU, US, China) referenced for comparative context
Japanese perspective on learning/imitation offered by Ken

Limitations & Notes

Transcript quality: Partial transcript with some unclear speaker attributions and incomplete sentences; some context may be lost.
Unfinished discussion: Panel appears to end mid-conversation as moderator notes "times up," and keynote by Nir Troy begins but is cut off.
Specific metrics: Production music statistic (42–48%) provided without source citation.
Generalization caution: While examples span music, film, gaming, and visual art, the panel does not systematically address all creative verticals equally.