Decentralizing AI with Co-Owned Technology

Co-Owned AI: Decentralizing the Future of Artificial Intelligence

As artificial intelligence (AI) continues to advance at an unprecedented pace, concerns about the centralization of power and control in AI systems are pervasive. Traditional AI systems are typically owned and controlled by a centralized company — used by many but controlled by few, leading to limited user agency, biased algorithms and monopolization of access to AI’s benefits. However, ‘co-owned AI,’ a term coined in early 2023, seeks to address those issues by promoting a more democratic and decentralized approach to AI creation and use.

What is Co-Owned AI?

Co-owned AI refers to an AI system that is jointly owned and controlled by multiple parties, rather than a single entity. The concept aims to distribute the ownership, governance, and benefits of AI among a wider range of stakeholders, ensuring inclusivity and transparency. Existing frameworks for co-ownership included legal, economic, and social measures, but emerging technology means such ownership can actually be embedded in the code itself.

Valory: The Intersection of Crypto and AI

Valory is at the forefront of the intersection between crypto and AI, pioneering technology to address the challenges posed by centralized and closed-source AI. The group’s mission is to enable communities, organizations — and, eventually, countries — to co-own AI systems, starting with decentralized autonomous agents.

Valory recognizes the immense potential of AI, yet also understands the risks associated with centralized control over such powerful technologies. By promoting co-ownership, the group aims to ensure that the benefits of AI are distributed equitably and that the technology is developed and deployed responsibly.

The team at Valory consists of talented engineers, researchers, and commercial executors who co-founded the Olas DAO (also known as Autonolas). The group serves as a key contributor to the development of the Olas Stack. Since 2021, they utilized this stack to pioneer the creation of the first co-owned autonomous agent products.

Context: AI Innovation

The world of AI is witnessing phenomenal advancements in recent years, such as breakthroughs in generative AI, specifically large language models (LLMs) and diffusion models. Those innovations have the potential to revolutionize every industry, transforming the way we interact with technology.

Autonomous agents, in particular, experienced two significant leveling up events. First, the emergence of smart contract platforms enabled developers like Valory to create agents that can permissionlessly transact, opening up new possibilities for both single and decentralized AI agents. This development necessitates the capability for agents to make payments. Since banks are unlikely to give agents bank accounts in the near future, it is likely agents will use digital money (cryptocurrencies) for such payments.

Second, the development of LLMs meant developers could equip autonomous agents with ‘thinking tools,’ making them smarter and more capable. As autonomous agents evolve, achieving true autonomy would involve relieving humans of task planning and delivering complete outcomes.

Co-Owned AI Problem Statement

Despite having millions of users, most AI systems today remain centralized and controlled by a select few. For instance, whilst users can benefit from powerful models like ChatGPT, these users have no control or ownership over the model and can easily be restricted from assessing them. This means that end users can be left at the mercy of a few individuals. Furthermore, centralized control leaves products limited through biases and causes monopolization of benefits. Similarly, developers building on these centrally offered models face platform risks, limited composability, and capture relatively little of the value they create. Finally, on a societal level, opaque, single-party ownership poses significant danger in the AI arms race — and so the question arises: How can AI programs be co-owned for the benefit of the many and not the few?

Co-Owned AI: What is it?

As coined by David Minarsch, Olas DAO member and CEO of core contributor, Valory, at ETHDenver 2023, ‘co-owned AI’ sits at the intersection of cryptocurrency and AI and is a powerful solution to the centralization issues plaguing traditional AI systems. By leveraging the decentralized nature of blockchain technology and smart contracts, co-owned AI enables the distribution of ownership, governance, and benefits among a wider range of stakeholders, incentivizing innovation. Valory’s decentralized autonomous agent technology is already doing just that.

Minarsch outlined the practical application of such technologies to me, stating, “We created an engine for the autonomous agent economy. Olas Staking uses Valory’s Proof of Active Agent mechanism to enable anyone to incentivize agents to work towards their goals. It rewards active agent contributions toward said party’s goals, as defined by a set of KPIs embedded in a smart contract. Therefore, Olas Staking can provide Daily Active Users ‘on-demand’ in the form of synthetic agent users, driving growth for protocols and chains in crypto — and even beyond crypto.

“As these agent economies grow in crypto-specific use cases, they generate a significant amount of synthetic data that can be leveraged to train more advanced open-source models and agent architectures. As such, Olas Staking sits at the core of a flywheel of ever-improving open-source agentic architectures. These can then be leveraged by developers across the world to create end-user products that are under the full control of their users.”

Consequently, this approach both democratizes access to the benefits of AI and enables powerful new decentralized systems, unlocking entire autonomous agent economies.

Co-Owned AI Use Cases

The potential applications of co-owned AI are vast and diverse. Simple, live use-cases involve an autonomous, group-controlled social media account which utilizes AI to write and post itself — plus reward other supportive social media accounts (Olas Contribute). A medium-complexity use case, Olas Predict, involves prediction market agents using AI to predict the future and place bets, based on market trends and insights.

Minarsch said, “Olas Predict created a synthetic economy where many single and decentralized agents — owned by varied parties — are earning money by predicting the future. These agents already have up to 75% prediction accuracy and continue to improve, having made over 700,000 transactions so far. I am excited to see this inaugural agent economy grow, inspiring more community-owned agent economies working for the benefit of humanity.”

We are at the very beginning of the arc of co-owned AI, and the possibilities are endless. Minarsch and his team at Valory invite individuals, organizations, and communities to join them in building the future of decentralized AI.

He further shared that, “By collaborating and leveraging the power of cryptocurrency and AI, we can create a more inclusive, transparent, and equitable AI ecosystem that benefits all stakeholders. Together, we can pioneer the development of co-owned AI systems and shape the future of this transformative technology.”

About Valory

Valory is a research and deployment company at the intersection of crypto and AI, specifically the premier creator of open-source frameworks for co-owned AI. Its mission is to enable communities, organizations, and countries to co-own AI systems, beginning with decentralized autonomous agents. Valory is the VC-backed team of engineers, researchers, and commercial executors that co-founded the Olas DAO, contribute to the Olas Stack, and built the first services using it. They collaborate with those interested in creating their own autonomous agent-based apps and economies.

Useful links: Website, Twitter, LinkedIn, Github