As generative AI reshapes industries from media and marketing to software development and healthcare, one question is becoming impossible for enterprises, policymakers and technology providers to ignore: Who should benefit when AI systems are trained on human-created content?
That debate has intensified as courts and regulators scrutinize how AI models are built, how synthetic media is distributed and whether creators deserve compensation when their work contributes to commercial AI products. Members of the Senior Executive AI Think Tank—a curated group of experts specializing in machine learning, generative AI and enterprise AI applications—say the future of AI depends on building sustainable systems that balance innovation with accountability, transparency and trust.
Lawsuits and copyright disputes over AI training data have accelerated globally, while companies such as Adobe continue advocating for licensed datasets and provenance frameworks designed to verify content authenticity. At the same time, enterprise adoption of generative AI continues to surge, with a McKinsey study on the state of AI finding that organizations are rapidly increasing investments in generative AI initiatives despite ongoing governance concerns.
The challenge now facing the industry is not simply whether AI companies should compensate creators, but how to build systems that make compensation, transparency and innovation sustainable at scale. Below, Think Tank members outline what that future could look like—from collective licensing models and provenance standards to creator opt-in frameworks, enterprise governance strategies and new approaches to trust in the age of generative AI.
Establishing Clear Legal Boundaries for AI Training
Kranthi Kumar Manchikanti of Microsoft argues that sustainable AI copyright policy requires multiple coordinated solutions rather than a single legal compromise. Microsoft has become deeply involved in AI licensing and content partnerships through its investments and integrations across generative AI systems.
Manchikanti says the industry needs “four separate fixes, not one grand bargain.”
He believes the first priority is drawing a clear legal distinction between lawful and unlawful sourcing practices. “Courts now agree: Training on lawful data is fair use; training on pirated data isn’t,” he says. “Codify that with real damages.”
He also points to provenance standards as a critical safeguard for high-risk content categories such as political media and news distribution. “Require C2PA Content Credentials for AI news, political content and large training sets,” he says. “Make ‘do not train’ tags legally enforceable.”
The Coalition for Content Provenance and Authenticity, known as C2PA, has gained support from major technology and media organizations seeking standardized verification systems for digital content authenticity.
Manchikanti additionally argues that the industry must separate content licensing for AI grounding from licensing for model training. “Grounding and citation deals are about distribution,” he says. “Training-data deals are about scarce archives. Don’t conflate them.”
For broad-scale creator compensation, he supports collective licensing frameworks similar to the music industry. “Pay creators ASCAP-style—blanket fees, sampled attribution—because per-output tracking isn’t technically possible,” he says.
Still, he offers a blunt assessment of the current market dynamics. “The training fight is mostly lost,” he says. “Creators get paid less than headlines suggest.”
Negotiating Accountability Into Enterprise AI Contracts
Lynn Comp, Head of AI Center of Excellence at Intel, says enterprises should approach AI licensing disputes through a commercial risk-management lens.
Intel, one of the world’s largest semiconductor manufacturers and a foundational force behind modern computing infrastructure, increasingly supports AI workloads across enterprise systems and data centers. Comp says organizations deploying frontier AI models must negotiate contracts carefully and understand their leverage.
“If the company is large enough to negotiate with frontier model companies, they should focus on the indemnification and residuals terms,” she says. “They should negotiate with explicit knowledge of their ‘walk away’ position.”
That strategy, she explains, places accountability on model developers regarding how training data was sourced and managed.
Comp also believes organizations building private AI systems should invest heavily in traceable and verifiable data pipelines. “For private models, adding into the supply chain data sources like companies with synthetic data capabilities and rigor around provenance for your own training data are critical,” she says.
“Regulation must avoid stifling innovation by creating barriers only hyperscalers can afford.”
Moving From Extraction to Partnership
Some Think Tank members emphasize that the future of AI depends on replacing extractive data practices with partnership-driven ecosystems.
Venkata Kondepati, Manager of Data Architecture and Engineering at Ascentt, says the industry must shift away from what he describes as a “scrape first, litigate later” mindset.
“A sustainable AI ecosystem needs a transparent value-exchange model,” he says. “The industry should adopt licensed training frameworks, standardized content provenance and usage-based compensation so creators share in the upside their work generates.”
Kondepati believes open standards will become essential for trust and traceability. “Open standards like C2PA can help establish trust and traceability,” he says.
Similarly, Aravind Nuthalapati, Cloud Technology Leader for Data and AI at Microsoft, argues that sustainable AI governance requires creators to participate directly in the economic value chain.
“The industry should adopt licensed and provenance-aware training pipelines, where creators can opt in, define usage rights and receive fair compensation,” he says.
Nuthalapati warns regulators against creating rules that only the largest AI companies can realistically navigate. “Regulation must avoid stifling innovation by creating barriers only hyperscalers can afford,” he says.
Both leaders see creator relationships as long-term strategic assets rather than compliance obligations.
“The long-term balance lies in treating creators as ecosystem partners—not raw data sources,” Nuthalapati says.
“The organizations that treat copyright, transparency and creator economics as core parts of the AI operating model—not legal afterthoughts—will lead the curve.”
Why Trust May Become AI’s Most Valuable Asset
Shreyas Nair, AI Operator for Wordsworth AI, believes the industry’s biggest risk is not simply legal exposure, but erosion of trust.
Nair, an AI founder-operator with experience in venture investing, enterprise consulting and commercialization, says organizations must stop treating creators as either “obstacles or free resources.”
“A sustainable approach does not mean training on everything or licensing every single pixel,” he says. “Instead, it should involve several layers: using premium datasets with clear licenses, giving creators ways to opt out or show they do not want their work used.”
He argues that cheap and untraceable training data is becoming unsustainable economically and reputationally.
“Leaders should realize that the days of cheap data are coming to an end,” Nair says. “They need to start building systems to track data sources, reduce legal risks, get creator consent and be transparent about synthetic media.”
If organizations fail to adapt, he warns the consequences may extend beyond courtrooms. “Copyright will become more than just a legal problem,” he says. “It will also affect trust and how widely people use their products.”
Richie Adetimehin, AI Advisory and Transformation Delivery Consultant at Visani America, agrees that trust will ultimately determine market leadership.
“The real long-term risk will be scaling AI without trust,” he says.
Adetimehin believes organizations should embed copyright governance directly into their AI operating models rather than treating it as a reactive legal function.
“The organizations that treat copyright, transparency and creator economics as core parts of the AI operating model—not legal afterthoughts—will lead the curve,” he says.
Building AI Compensation Systems That Scale
Sidhesh Badrinarayan, Engineering Lead at Google, believes the industry needs a compensation system modeled after digital music platforms.
“AI training is stuck in a false choice: Scrape everything for free, or pay millions in upfront licensing,” he says.
Instead, Badrinarayan proposes “a Spotify model for AI.”
“We should pay creators micro-royalties based on actual use,” he says.
His framework would compensate creators dynamically whenever AI systems rely on their style, code or content to generate outputs. “The moment it uses a specific creator’s style, code or article to generate an answer, a micro-payment is triggered,” he says.
To support that infrastructure, Badrinarayan advocates for automated watermarking and identification systems. “Just like Shazam identifies a song, built-in digital watermarks can instantly track which creator’s work is being used, making payouts automatic and transparent,” he says.
Andre Shojaie, Founder of HumanLearn, similarly believes sustainable AI compensation requires systems-level thinking rather than isolated policy fixes.
“Sustainable AI copyright is less a legal question than a design challenge,” Shojaie says.
He envisions ecosystems where data provenance, attribution and creator compensation are built directly into AI infrastructure. “Every piece of training data carries its origin like a root,” he says, “and every model output traces back to the creators who fed it.”
For Shojaie, future AI governance systems must function almost invisibly in the background. “Compensation flows automatically, invisible but real,” he says.
“If creators, publishers, artists and researchers become unpaid infrastructure for AI systems, the industry eventually loses trust and quality.”
Compliance as a Competitive Advantage
Sabarinath Yada, Business Architect Associate Manager at Accenture, says enterprises should begin treating AI copyright governance as a core operational discipline.
Accenture advises organizations globally on digital transformation, AI modernization and enterprise technology strategy, giving Yada direct visibility into how governance concerns shape enterprise adoption decisions.
“Industry can no longer treat copyright compliance as a secondary, post-deployment consideration,” Yada says.
He argues that provenance and creator compensation should receive the same level of operational rigor as cybersecurity or performance engineering. “Compliance from day one should be the baseline, not an afterthought,” he says.
Yada believes the market will increasingly reward trusted and legally verified data ecosystems.
“A sustainable AI ecosystem will increasingly favor verified, tiered data marketplaces where high-value, legally cleared datasets command premium pricing,” he says.
For enterprises, that may create both legal protection and strategic differentiation. “Compliance becomes a competitive differentiator enabling ethical, high-velocity innovation,” Yada says.
Independent researcher Aishwarya Shah also believes governance and trust will increasingly define market leadership.
“The companies that lead responsibly will treat governance and creator trust as competitive advantages, not compliance burdens,” Shah says.
She warns that invisible extraction models may ultimately damage the quality and sustainability of AI itself. “If creators, publishers, artists and researchers become unpaid infrastructure for AI systems, the industry eventually loses trust and quality,” she says.
Strategies for Navigating AI Copyright and Trust
- Create clear AI data governance standards internally. Organizations should establish formal policies around provenance, licensing and acceptable training data sources before deploying generative AI systems.
- Negotiate indemnification aggressively. Enterprise leaders should ensure AI vendors clearly assume responsibility for training-data compliance and intellectual property risk.
- Treat creators as ecosystem partners. Sustainable AI systems will increasingly depend on opt-in licensing and shared economic value rather than extractive data collection.
- Invest in opt-in licensing and creator consent systems. Move beyond low-cost data acquisition strategies by building mechanisms for creator approval, source tracking and synthetic media transparency.
- Build trust into AI products from the start. Transparency around synthetic media, training data and creator rights can become a competitive differentiator.
- Explore scalable royalty and collective licensing models. Music-industry frameworks may provide useful templates for compensating creators at AI scale.
- Design AI ecosystems with traceability built in. Think beyond compliance and create systems where attribution, provenance and compensation are embedded directly into AI infrastructure.
- Embed copyright governance into enterprise architecture. Provenance and creator compensation should be treated with the same rigor as cybersecurity and compliance.
- Turn responsible AI governance into a market advantage. Organizations prioritizing creator trust, attribution standards and transparent governance may strengthen both AI quality and long-term customer confidence.
- Prepare for premium licensed-data marketplaces. Legally verified datasets may become strategic assets that improve both AI performance and enterprise legal protection.
The Next Chapter of AI Will Be Built on Trust
The AI industry is approaching a moment that feels a lot like the early days of streaming music or social media—an inflection point where explosive growth is colliding with questions nobody fully answered before the technology went mainstream. The difference is that this time, the stakes extend far beyond entertainment. The data powering AI systems increasingly shapes business decisions, public trust, creative industries and even political discourse. That makes the fight over copyright and compensation about more than ownership. It is about defining the rules of participation in the next digital economy.
What emerges from the perspectives shared by members of the Senior Executive AI Think Tank is not a call to slow AI down, but a recognition that speed without trust eventually creates friction. The companies most likely to lead the next era of AI may not simply be the ones building the largest models, but the ones building systems people believe in—systems where creators are visible instead of invisible, provenance is expected instead of optional and innovation does not come at the expense of accountability.
