Fair AI is the only viable solution to combat the growing risks of Big AI. As we speak, the world’s largest tech companies are building powerful AI-enabled applications. Of course, AI is laden with economic opportunity; but it presents looming existential threats as well. The market is dominated by companies like Google, OpenAI, Microsoft, Meta, and Nvidia — many of whom have long histories of user data exploitation.
These companies will benefit financially from the vast troves of user data that will power their LLMs. The end users — you and me — won’t see a dime of that financial gain. As AI replaces everyday jobs at an accelerated pace, users being compensated for their data contributions will grow more and more important.
The financial opportunity of AI needs to be distributed in a way that is available, balanced, and equitable — but still competitive. In other words, it’s time for Fair AI.
Big AI is exclusive and exploitative
At its core, AI is a data play. Whoever owns the data owns the future of AI. For the last 25 years, the world’s biggest tech companies have sourced our data, largely free to use it as they wish with vague terms and conditions conveniently dangled in front of us just before the dopamine hit of a product purchase. Big Tech is now transforming into Big AI.
AI has now placed a premium on data. More valuable than ever, the gathering, scraping, and exploitation of users’ data by these tech behemoths will only accelerate. Just recently, Adobe faced global backlash after instituting “Terms and Conditions,” giving them the rights to use all creative work from users of their products, and are now urgently backpedaling (PCMag). Importantly, Adobe did not have plans to compensate a creator if their work was used by an LLM to inform a generative AI output.
This world — in which data’s financial opportunities incentivizes companies to hoard more and more without an eye towards user privacy or creative sovereignty — is undeniably a dangerous one. It’s a world in which all the issues of today’s tech monopolies are amplified a hundred times over under the economic forces of AI and data.
The knee-jerk reaction might be to “turn off the faucet” — to deny any company from ever using user data. But the future is not one in which user data will never be used. AI is inevitable, and it is imperative that instead of stopping AI, we redirect its course towards something more sustainable, equitable, user-centered, and properly incentivized.
The data that AI is trained on comes from the millions of people interacting daily with online applications and services. This data should be monetized by those contributors — in other words, by you. We call this Fair AI — and we believe it is not just a good thing to have in AI; we believe it is fundamental to ensuring that as AI transforms our society, it does so to the benefit of everyone.
The world needs fair AI
Fair AI is the amalgamation of technologies and economic incentives that, together, ensures AI evolves in a way that is beneficial to everyone who enables or uses it. There are some core principles to Fair AI — ownership, permission, and fair compensation. Specifically, fair compensation for the contribution of data, compute, and content to datasets. Consider this example -- a user could give permission to contribute their data from their social profile – say, Twitter/X. This is not dissimilar from the process that happens when you give certain applications access to your calendar. These contributions create new datasets for developers to use that belong outside of what's publicly available and because of this, that person contributing their data to then build a more robust AI ecosystem, would then receive compensation in the form of an on-chain asset. This type of contribution is distinctly different from the current direction of Big AI. You are not compensated when the decades of your Google Search activity is used by Gemini to inform a generative output for another service.
Decentralized technology is critical to achieving Fair AI. Only through an open and secure network like a blockchain can we ensure that user data can be permissioned, tracked, revoked, and yet still leveraged by powerful LLMs. When you take into account the native incentivization of blockchain networks, you introduce the most powerful and resilient feature of Fair AI: returning value to the people responsible for contributing that data in the first place. Apps are already being built on top of blockchain that enable users to contribute their data to decentralized LLMs and earn in the process.
There are many blockchain solutions that are powering decentralized AI as momentum picks up, as highlighted by VanEck's recent report stating that the global Decentralized AI market is expected to grow to $10.2 billion by 2030. Crypto solutions like on-chain data markets, decentralized identity, zero-knowledge proofs, and on-chain AI governance are driving the projected growth of this space, reflecting the rising demand for decentralized systems that provide users with greater economic benefits for their contributions. When implemented, these solutions help create a more equitable and fair AI economy.
Fair AI, however, does not just make today’s AI more equitable. It also makes it more powerful for the AI era. Unlike data sourced from closed apps or held on closed LLMs, data on open networks is not separated by silos. Any data that has been made available by the contributor can be leveraged, meaning AI developers can theoretically access even greater troves of data to train LLMs. Effectively, the world’s data is available to them, rather than just the dataset from a community of app users.
The permissionless nature of blockchains also fuels more AI innovation at a global level, because any developer can tap into any permissionless dataset. Lastly, the monetization element of Fair AI incentivizes the creation of higher-quality, higher-value, real-time datasets. If people know that certain data is valued more by the market, they will endeavor to deliver that data to the open market — ultimately creating a more robust, complete, balanced data ecosystem.
AI is inevitable, but its exploitation does not have to be. Fair AI is the only viable, realistic pathway for us to continue innovating AI while curtailing the more sinister risks it poses when it remains in the hands of a few tech giants. By supporting Fair AI and those pursuing it, we can create a future not just where the economic benefits of AI advancement are accessible to everyone, but where the AI solutions are even more powerful than those we imagine today.
Note: The views expressed in this column are those of the author and do not necessarily reflect those of CoinDesk, Inc. or its owners and affiliates.
