My Background and Motivation
My research specialization focuses on the socioeconomic impacts of AI, with particular attention to causality problems and solutions that yield mutual benefit - ensuring AI adoption doesn't come at the expense of human wellbeing, social cohesion, and economic opportunity.
My work spans three critical domains: AI operations (I won an AI hacking bounty breaking Meta's LLMs and have trained image generation models, FinTech (having interned at Chainlink Labs and contributed to adding payment rails to Google's A2A protocol), and socio-economics (my academic background). This intersection gives me a unique vantage point on both the technical possibilities and human costs of AI development.
What drives this research is both deeply personal. I'm surrounded by people working in AI automation, but I also have many more friends outside this sphere in knowledge-based roles. The reality is stark: I can automate the tasks that keep them employed, or I could name the person or company currently building their replacement. Rather than accepting the false choice between technological progress and human displacement, I believe we're at a pivotal moment where we can reshape how expertise is valued and compensated in the AI economy.
This concern was crystallized when I heard Geoffrey Hinton, a pioneer of early AI development, on the DOAC podcast. His core recommendation to people facing AI displacement was to "become a plumber," but he offered no substantive solutions to the externalities created by AI labor transitions. While he suggested UBI as a potential remedy, he acknowledged such strategies are fundamentally flawed. This highlighted a crucial opportunity: rather than treating human labor empowerment as an afterthought, we can build AI development from the ground up to create genuine partnerships between human expertise and artificial intelligence.
Correcting the AI Industry's Greatest Inequity
The AI industry has inadvertently created one of the most glaring compensation disparities in modern technology, but this presents an unprecedented opportunity for change. Internal contributors (AI scientists training models, MLOps engineers deploying pipelines) typically receive restricted stock units (RSUs) that give them equity stakes in the companies they help build. Their compensation grows with the long-term success of the AI systems they develop.
Meanwhile, external contributors face an entirely different reality that's ripe for transformation. Annotators, evaluators, and domain experts working on contract basis receive only wage compensation, despite providing equally essential contributions to AI system development. This disparity isn't accidental, it's structural, rooted in outdated contract-driven approaches that haven't evolved with the reality of AI development.
But here's the opportunity: this two-tiered system is so obviously inequitable that it creates a powerful case for change. As AI companies capture enormous value, the professionals providing irreplaceable domain expertise are positioned to demand their fair share. The question isn't whether this system should change, but who will lead that transformation.
The Fleeting Nature of AI Labor Opportunities
The current AI landscape presents a complex picture of labor displacement and opportunity. While basic data annotation work - particularly in developing countries - faces increasing automation through synthetic data generation and AI-driven evaluation, a new tier of specialized work has emerged.
The Exploitation of Basic Annotators
The treatment of basic data annotators, particularly in the Global South, reveals the exploitative potential of current AI labor practices. A recent DW Documentary shed light on the reality of data workers, exposing the hidden workforce behind AI systems. These workers evaluate data for AI models on content ranging from everyday interactions to disturbing material, often for minimal compensation and with severe psychological costs. The documentary reveals how overnight, work can simply disappear as companies shift to automated alternatives, leaving workers with no recourse or transition support.
This pattern of exploitation - extracting value from human labor while providing minimal compensation and no stake in the resulting systems - represents exactly what we must avoid as AI capabilities expand into more specialized domains.
The Rise of Expert-Led AI Development
As large language models become more powerful and expand into specialized fields like medicine, law, and scientific research, we're witnessing the emergence of a new paradigm: expert-led AI development. The demand for subject matter experts has skyrocketed because high-performing AI systems require the kind of nuanced judgment that only domain specialists can provide.
This represents far more than a market shift, it's the foundation for a fundamental restructuring of how AI systems are built. Domain experts are no longer peripheral contributors to AI development; they're becoming central architects of how AI operates in their fields. This transformation creates unprecedented leverage for professionals who recognize their strategic importance.
Historical Precedent: The 1980s Productivity-Wage Divergence
This moment echoes a critical inflection point in economic history. The productivity-wage divergence that began in the 1980s offers a stark lesson about what happens when technological advancement concentrates benefits among capital owners rather than workers.
From 1979 to 2019, several key mechanisms drove this divergence:
Erosion of Worker Bargaining Power: Union coverage fell from 27.0% to just 11.6%, representing a massive decline in workers' ability to capture productivity gains through wage negotiations. Union membership decreased from nearly one in three workers in the 1950s to less than one in ten by 2024.
Technological Change and Capital Intensity: Advances in automation and technology increased productivity while potentially reducing demand for certain types of labor, allowing employers to capture more productivity gains rather than sharing them with workers.
Globalization and Market Concentration: Increased competition from lower-wage countries and the concentration of economic power in fewer, larger firms reduced worker leverage and allowed companies to suppress wage growth even as productivity rose.
Shift in Corporate Governance: The rise of shareholder primacy since the 1970s led companies to prioritize returns to capital over sharing productivity gains with workers.
The result was that productivity improvements flowed disproportionately to capital owners and executives rather than being shared with the workers who helped generate them.
We're at a crossroads where professionals can choose to lead transformation rather than accept displacement. Rather than selling services for short-term gain at the cost of long-term economic participation, domain experts can pioneer new models where specialized contributors receive equity as recognition of their essential role in AI development.
The traditional model of shareholder primacy that drove the 1980s divergence stands in stark contrast to the collective ownership alternatives that forward-thinking industries are already implementing. Worker cooperatives, employee stock ownership plans, and profit-sharing arrangements demonstrate that productivity gains can be distributed equitably while driving innovation. The question for AI development isn't whether these models work, but which professionals will be bold enough to lead their implementation.
The Strategic Advantage of Domain Experts
Far from being passive participants in AI development, domain experts possess unique characteristics that position them as natural leaders in creating more equitable compensation models:
Intellectual Capital as Power
Domain experts' hyper-specialized skillsets make them natural leaders in professional collective action, unlike generic tasks that can be taught quickly. Their qualifications and expertise command premium value in AI system evaluation, particularly when providing assurance reviews for high-stakes applications where mistakes have real-world consequences.
Most importantly, these professionals have made substantial investments in their expertise: years of post-graduate education, decades of practice, and ongoing professional development. This intellectual capital represents genuine power in negotiations. Unlike basic annotation work, domain expertise cannot be easily replicated or outsourced, giving these professionals leverage that many other workers lack.
This investment profile makes domain experts natural equity partners rather than wage workers. Professionals who have invested years building specialized knowledge deserve compensation models that provide returns commensurate with their investment, just like other forms of capital. They're not asking for charity; they're demanding recognition of their true market value.
Feasible Tokenization
The specialization of subject matter makes defining tokenizable contribution categories more feasible than with generic tasks. This creates the foundation for new economic models that can track and reward contributions across multiple companies and projects.
Building Professional Guilds for the AI Age
When organized through modern professional guild structures, domain experts can unlock transformative opportunities that go far beyond traditional employment:
- Creating domain-specific benchmarks and standards that become industry requirements, positioning experts as regulatory leaders rather than service providers
- Governing market pricing through collective bargaining power, ensuring compensation reflects true value contribution
- Driving continued education and research on how AI should be adapted for their specific fields, making them innovation leaders
- Taking on larger projects directly from AI labs that would typically require intermediary data annotation companies, capturing more value in the process
These capabilities transform domain expert collectives from cost centers into strategic partners and competitive differentiators for AI companies. Organizations that can access high-quality domain expertise through these new models will offer premium services, stronger quality assurance, and more reliable AI systems, creating genuine competitive advantages in increasingly crowded AI markets.
Rather than competing individually for contracts, organized domain experts become indispensable partners in shaping how AI develops in their industries.
Pioneering the Future of Work Through Technology
My ongoing research focuses on developing the protocols and tokenomics that can make these new models of participation a reality. This isn't just about fairness, it's about building the infrastructure for the future of expert work.
The technical approach leverages both blockchain technologies and MLOps systems to create something unprecedented:
Blockchain enables: Revolutionary micropayment systems and smart contracts that can programmatically distribute tokenized equity to diverse stakeholders across complex, multi-step contribution processes. This solves the transaction cost problems that make tracking individual contributions infeasible for generic work.
MLOps integration ensures: These protocols integrate seamlessly with current AI development practices, creating market-driven opportunities for domain experts without requiring wholesale industry restructuring.
This approach represents a modern evolution of collective ownership, where specialized contributors can maintain equity stakes across multiple projects and companies. Instead of the traditional model where only internal employees benefit from long-term value creation, tokenized systems can distribute ownership more broadly while maintaining the efficiency and flexibility that modern AI development demands.
The goal is creating platforms that enable domain expert collectives to participate directly in AI development and production, particularly in training and evaluation phases where their expertise provides irreplaceable value. These platforms create genuine business value by offering companies access to higher-quality expertise and more reliable AI systems, making the business case for more equitable compensation models compelling rather than charitable.
Seizing the Moment of Maximum Leverage
The professionals who act now will shape the future of work in the AI age. We're in a brief but crucial window where domain experts have maximum leverage. As AI systems become more capable, the current reliance on domain expert evaluation will evolve, but not disappear. The experts who organize now, who claim equity stakes rather than accepting only wage compensation, will be positioned to lead and benefit from the continued development of AI systems in their domains.
This moment represents more than avoiding displacement, it's about claiming leadership in defining how AI develops in specialized fields. For professionals who have invested years in education and decades building expertise, this represents the opportunity to ensure their intellectual capital investments generate the long-term returns they deserve.
The choice facing domain experts today is transformative: repeat the pattern of the 1980s (selling their expertise for immediate compensation while the long-term value accrues to capital owners) or pioneer new models where expertise commands equity. It's a choice between accepting short-term payments for long-term value creation, or leading the development of compensation models that recognize the true nature of expert contributions to AI systems.
Domain experts who act now won't just protect their economic interests, they'll become the architects of more equitable and effective AI development practices.
The question isn't whether AI will transform these professions, but whether the professionals will lead that transformation or be swept along by it.
Leading the Transformation
This research agenda represents more than an academic exercise: it's a framework for empowering domain experts to lead the development of AI systems that genuinely serve human expertise rather than replacing it. By creating mechanisms for equitable participation in AI value creation, we're building a future where technological advancement amplifies human capability rather than diminishing it.
Domain experts have a unique opportunity to be pioneers in this transformation. They possess the expertise, the leverage, and increasingly, the technological tools to reshape how specialized knowledge is valued in the AI economy. The question isn't whether change will come, but who will lead it.
The window is open, but it requires bold action. The time for domain experts to organize, to demand equity rather than wages, and to claim their role as leaders in AI system development is now. Those who act will define the future of expert work in the AI age.
All the best,
Beau