Disclosure: The views and opinions expressed here belong solely to the author and do not represent the views and opinions of crypto.news’ editorial.
All industries are becoming more reliant on AI to support day-to-day operations. Even in the crypto space, AI has been a driver for adoption. However, underneath the surface, the mechanics that power an AI are severely flawed, creating bias and discrimination in its decision-making. Left unattended, this will limit the potential of the technology and undermine its purpose in key markets.
Summary
- Regulatory action on ethical AI has stalled, leaving it to the industry to self‑police data sourcing, annotation, and fairness — or risk compounding systemic bias.
- Blockchain‑based, decentralized data labelling offers both transparency and fair compensation, especially for underrepresented contributors and emerging economies.
- Stablecoin payments ensure equitable rewards globally, transforming data annotation into a viable income stream capable of rivaling local living wages.
- In the AI arms race, better data means better performance, and decentralization turns diversity from a moral obligation into a competitive edge.
The solution to this challenge lies on the blockchain. Leveraging the same decentralized technology that enables greater transparency in transactions can also enable increased fairness in how AI is built and works.
The source of bias
AI’s bias stems from the underlying data that is used to inform the technology. This data — which can include everything from audio clips to written content — needs to be ‘labelled’ for the AI to understand and process the information. However, studies have shown that up to 38% of data could hold biases that may reinforce stereotypes based on gender or race.
More recent research continues to confirm the problem. For example, a 2024 study of facial expression recognition models found that Anger was misclassified as Disgust 2.1 times more often in Black females than in White females. Additionally, a 2019 NIST benchmark review determined that many commercial facial recognition algorithms inaccurately identified Black or Asian faces 10 to 100 times more frequently than white faces, highlighting how skewed datasets lead to disproportionately higher error rates for underrepresented groups.
It’s here that discussions around ‘ethically’ using AI often come to the fore. Unfortunately, this topic is being deprioritised through regulation and the perceived belief that an ethical approach to AI will limit profitability. This ultimately means that ethically sourcing and labelling AI data is unlikely to come from governments anytime soon. The sector has to police itself if it hopes to establish longstanding reliability.
Decentralizing the data sourcing
Overcoming AI bias requires sourcing ‘frontier data’: high-quality, diverse datasets created by real individuals from underrepresented communities, which can capture the nuances that legacy datasets consistently miss. By engaging contributors from varied backgrounds, the resulting datasets become not only more inclusive but also more accurate. Blockchain offers a powerful tool in advancing this approach.
Integrating blockchain into a decentralized data annotation process helps enable and validate fair compensation for contributors. It brings full traceability to every data input, allowing for clear attribution, better oversight of data flows, and stricter controls based on the sensitivity of a given project. This transparency ensures that data is ethically sourced, auditable, and aligned with regulatory standards, addressing long-standing issues of exploitation, inconsistency, and opacity in traditional AI data pipelines.
Creating opportunities
The opportunity goes beyond fairness, as blockchain-based labelling also creates powerful growth potential for emerging economies. By 2028, the global data annotation market is expected to reach $8.22 billion. Yet even this may underestimate the sector’s true potential, given the rapid proliferation of AI technologies, the underwhelming performance of synthetic training data, and the increasing demand for high-quality training data. For early adopters, particularly in regions with limited existing infrastructure, this presents a rare opportunity to shape a critical layer of the AI economy while generating meaningful economic returns.
Debates continue to rage about AI stealing jobs from human workers, with some speculating that as many as 800 million jobs could be lost. At the same time, enterprises will increasingly prioritize robust datasets to ensure AI tools outperform human employees, creating a new space for individuals to earn income through data labelling and enabling the rise of new regional powerhouses in this service sector.
A stable return
Using the blockchain in AI labelling goes beyond payment transparency. Leveraging a consistent asset, such as a stablecoin, means that users will be fairly compensated regardless of their location.
All too often, manual-intensive roles have been outsourced to emerging markets, with companies undercutting one another to receive business. While legacy processes may hold back established sectors like manufacturing and farming, the emerging landscape of AI labelling doesn’t need to fall victim to this unfair practice. A stablecoin payment system ultimately means equality across markets, empowering emerging economies with an income stream that can rival their national living wage.
Profitable and equitable
Those with the best data will have the best AI. Just as financial markets once competed to the millisecond for faster internet connections, where even tiny delays translated into millions in gains or losses, AI now depends on the quality of its training data. Even modest improvements in accuracy can drive massive performance and economic advantages at scale, making diverse, decentralized datasets the next critical battleground in the AI supply chain. Data is where the convergence of web2 and web3 can have one of its biggest and most immediate impacts, not through displacing legacy systems, but by complementing and enhancing them.
Web3 is not expected to replace web2, but to become successful, it must fully embrace integration with existing infrastructure. Blockchain technology offers a powerful layer to enhance data transparency, traceability, and attribution, ensuring not only data quality but also fair compensation for those who contribute to its creation. It’s a common misconception that an ethics-led business cannot also be profitable. In today’s AI race, the demand for better, more representative data creates a commercial imperative to source from diverse communities around the world. Diversity is no longer a checkbox; it’s a competitive advantage.
Even as legislation lags or deprioritises ethics in AI, the industry has a chance to set its own standards. With frontier data at the core, AI companies can not only ensure fairness and compliance but also unlock new economic opportunities for communities, contributing to the future of intelligent technologies.
