As Congress conducts hearings with SEC Chairman Gary Gensler and pushes to better regulate and stimulate the evolving digital economy, it must recognize the unique needs of decentralized AI – a critical yet often overlooked sector at the intersection of blockchain and artificial intelligence.
Despite this crossover, decentralized AI cannot be legislated through a financial lens, nor can it be forced into AI regulations. Given its overlap in distinct sectors, however, there’s a real chance lawmakers will try to fold it into AI and crypto bills – or overlook it altogether – which would be a missed opportunity for innovation in this country.
Simply put, decentralized AI allows for the distribution of data, computation, and decision-making processes across multiple devices or nodes, enabling them to work together without relying on a centralized authority, often utilizing open source software and models. This gives developers the tools to share their data collaboratively to build AI models, and to access compute from a diverse range of sources. It’s a powerful framework that empowers these developers to contribute to the AI ecosystem without the need to manage the entire process themselves, enabling researchers and startups to participate in a field where rising costs and difficulty of access threaten to push them out.
That’s why it’s imperative lawmakers not overlook decentralized AI as they begin to regulate AI. It’s probably human nature to ignore, considering the broader AI industry is exploding and dominated by some of the world’s biggest corporations. They’re acquiring startups, pushing advancements, and launching new products at a breakneck pace. While there’s nothing wrong with Microsoft, Meta, Alphabet, and others investing heavily in the industry, lawmakers need to create space for researchers, entrepreneurs, and developers to thrive as well; that entails among other things remaining vigilant on antitrust, and ensuring that government-backed R&D funds don’t benefit the giants alone.
And while Congress is making welcome progress to advance comprehensive legislation that would clarify the rules of the road for crypto, the bills are one-dimensional – financial in nature – and don’t address the obvious difference between an underlying digital asset of a protocol versus the decentralized AI applications running on the same protocol. Financial regulators shouldn’t end up overseeing decentralized AI just because projects issue tokens; that would be like the SEC regulating toothpaste because Johnson & Johnson issues common stock.
It’s important we get this right, because decentralized AI is a critical field that can salvage the organizations that gave birth to the AI revolution – and possibly bring us the next wave. The universities that created the concept of machine learning and neural networks are under threat, as they cannot compete with Big Tech as it snaps up the world’s GPUs and top talent. In a similar vein, the kinds of startups that brought AI to market are facing similar constraints and oftentimes must defer projects. Without those engines of U.S. innovation working properly, progress will rest in the hands of a few large corporations. It’s a lesson that lawmakers will want to heed, as these universities and small businesses help form the fabric of the American economy, and create jobs and opportunities for constituents.
Decentralized AI can counter this trend of market concentration. The field allows organizations with limited funds to share data and compute across nodes, with developers stitching together a network of multiple databases to feed their models and dispersed GPUs to power them. It’s a more affordable solution that allows smaller players to participate and contribute. It also addresses many of the concerns about data privacy: developers can process data locally, which minimizes the need to transfer sensitive information to centralized servers. This reduces the risk of data breaches and unauthorized access.
All this goes to show the importance of establishing rules and regulations that allow decentralized AI to flourish. Congress must establish clear regulations for data privacy and security to protect individuals’ information when processed locally, and promote transparency and accountability in AI algorithms that can protect ethical standards and prevent bias in decentralized systems.
Lawmakers should consider a unique regulatory framework for decentralized AI, separate from traditional AI regulations in the Algorithmic Accountability Act and from financial guidelines covered in the Blockchain Regulatory Certainty Act. The new legislation would need to address the particular challenges of decentralized AI, ensuring transparency in algorithmic decisions, and providing guidelines on accountability in decentralized networks. Further, Congress should look at expanding funding opportunities under the National AI Initiative Act to ensure decentralized AI projects – especially those from universities and small businesses – receive government R&D funds and tax incentives, preventing Big Tech from monopolizing AI innovation. Along those lines, Congress would be wise to consider bolstering antitrust enforcement to stop major corporations from controlling critical AI infrastructure.
And the responsibility doesn’t rest solely on Congress’ shoulders. With lawmakers paying renewed attention to crypto and AI, now is the time for proponents of decentralized AI to speak up and advocate for the industry. And let’s face it, the industry hasn’t always done a stellar job of communicating what it does and what it needs to thrive. Yes, the technology is complex, but that doesn’t mean advocates can’t go beyond the technical language and articulate it in a way staffers and legislators without engineering degrees can understand. And most importantly, it’s critical to underscore the importance of getting this right, as a way to preserve innovation, jobs and opportunity; that’s the kind of language that resonates with members of Congress, and will advance this field to the next level.
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.
