AI and Crypto: Backbone of the Future Digital Economy
Disclaimer: The following article is intended to provide an overview of some aspects of web3, a term that refers to the decentralised and distributed applications that run on blockchain technology. The article does not constitute investment advice, nor does it endorse or recommend any specific cryptocurrency, project, or platform. Cryptocurrencies are highly volatile and risky assets that may lose value rapidly and unpredictably. You should do your own research and consult a professional financial advisor before making any investment decisions involving cryptocurrencies. The information shared in this article is purely for educational purposes and does not reflect the opinions or views of the author or any affiliated organization.
AI has taken the world by storm, specifically chatGPT by OpenAI, with its rapid adoption and potential to reshape technology interaction. The launch of its consumer app has seen developers, writers and other knowledge workers reporting on productivity boosts anywhere from 50% to 500%. This leads us to an interesting question: Could AI and cryptocurrency combine to create a new, transformative synergy in the future? In this evolving tech landscape, AI is akin to the arrival of the internet, while cryptocurrency signifies a radical shift in digital economics. Their merger could be a game-changer.
The impact of AI has been the talk of the town, and where it takes us in the future is an exciting but scary thought exercise to say the least, but nobody is denying the potential exponential increase in efficiency it can bring to society. What about crypto? Recently, Jason Choi of Tangent Ventures put out a question on Twitter asking his followers on how does Web3 increase societies productivity, and I agree with a reply by @ryanberckmans stating that Web3/Crypto ultimately reduces aggregate transaction costs to the digital economy. The TLDR: Web3 significantly lowers transaction costs by enabling secure data and money transfers on a global scale. Its decentralized nature minimizes risk, further reducing costs. Thus, it democratizes the global trade network, benefiting everyone, not just those in wealthy countries.
That said, I believe the future would be that AI emerges as the cerebral powerhouse of technology and that crypto/web3 lays down the economic rails, guiding the transactions (information or financial) interactions of a digital economy.
Use Cases
This blend could open up new possibilities that we haven't seen before, impacting everything from data security to problem-solving. Note that I am not an expert on AI by any means nor am I an expert on cryptography and crypto economics. This is a purely speculative take into what may happen in the future with AI and crypto. That said, let’s dive into the thesis.
The AI Native Reserve Currency
The digital economy, increasingly dominated by artificial intelligence, holds the promise of a universal reserve currency. This idea suggests a currency that AI systems across the globe could use for transactions, much like how fiat currencies function within traditional economies. It's an intriguing concept that could facilitate seamless AI-to-AI transactions and standardize a global AI economy.
Existing cryptocurrencies like Bitcoin (BTC) or Ethereum (ETH) might initially appear as suitable candidates for this role, given their established networks and recognition. However, they present certain limitations. Bitcoin, despite its trailblazing status, lacks advanced privacy features that could be crucial for preserving transactional confidentiality in an AI-driven economy. Ethereum, while boasting smart contract functionalities, grapples with scalability concerns which may hinder its utility in a global AI economy.
In contrast, privacy-focused cryptocurrencies offer an appealing alternative, given their emphasis on transactional security and anonymity. Additionally, the advent of newer projects such as World Coin opens up exciting prospects. These currencies could be specifically tailored to meet the distinct needs of AI systems, addressing the unique challenges posed by an AI-centric economic framework.
Ultimately, the ideal reserve currency for an AI-dominated world might still be on the drawing board, yet to be created. It would need to balance privacy, scalability, security, and global acceptance, a combination that the current crypto landscape has yet to fully realize. As we delve deeper into this realm, the perfect candidate might just be around the corner, waiting to be discovered. The future native currency of the digital currency may or not be an existing crypto, we would just have to wait and see.
AI Training
Training AI models, especially large language models (LLMs) like GPT-4 or other resource-intensive tasks, require substantial computational power. This requirement typically necessitates significant investment in data centres and high-performance computing infrastructure, often centralizing AI training in entities with sufficient resources. However, the convergence of AI and blockchain technology could shift this dynamic by decentralizing the process and making it more accessible.
Incentivizing Participation:
As companies like NVIDIA continue to push the boundaries of computing power with optimized GPUs designed for AI, the potential for advanced AI training increases. However, these powerful resources are often underutilized. Through blockchain networks, these optimized computing resources can be more effectively leveraged.
Individuals and organizations with such high-performance hardwares can contribute to a decentralized cloud, lending their processing power for AI training. In return, they can receive crypto as compensation. This process ensures the optimal usage and allocation of computational power towards the development of AI and incentivizes further R&D in AI-optimized hardware.
This incentive structure, automated and enforced through smart contracts, provides transparency and reliability in transactions. As a result, a symbiotic relationship between hardware advancement and decentralized AI training can be established, where the progress in one area fuels growth in the other. This could lead to an efficient and collaborative ecosystem that promotes the development and democratization of AI in a cost-effective manner.
Democratizing AI Training, Broadening Perspectives and Enhancing Accuracy:
Decentralized networks and cryptocurrency incentives not only democratize AI training and promote diversity, but they also open the door for the inclusion of subject matter experts who can provide unique insights and accurate data.
For instance, a cultural ethics expert from an emerging country could participate in the AI training process, providing nuanced perspectives and knowledge to guide the model's understanding of that particular culture. This could help avoid potential bias, misunderstandings, or oversimplifications in the AI's output related to that culture.
Similarly, in a business context, a go-to-market specialist from a specific region, such as Southeast Asia, could contribute their expertise to an AI model focused on business strategies. Their insights about the unique dynamics, challenges, and opportunities in these markets can improve the AI's ability to provide effective and contextually relevant business strategies.
By involving experts in the training process, we can further increase the accuracy, relevance, and depth of the AI's output. This results in AI models that are not only more representative and inclusive but also more precise and knowledgeable, ultimately leading to more effective and useful AI tools for a wide range of applications.
Enhancing Privacy: Don’t trust, verify.
Blockchain technology, with its decentralized nature, can significantly enhance the security and privacy of AI operations. The data used for training can be encrypted and securely shared across the network, preventing any single participant from having complete access to the data. Furthermore, it can also contribute to trust minimization when proprietary Machine Learning (ML) models are offered as a service. Mohamed Fouda from Alliance DAO explains this elegantly in his recent post.
A specific application of this is Zero-Knowledge Proofs (ZKPs), a cryptographic technique that allows one party to prove to another that they know a value, without sharing any other information aside from the fact they know it. ZKPs can be used to assure users that they are receiving the service they are paying for, and that the same model is applied fairly to all users1.
In this architecture, the ML model creator generates a zero-knowledge circuit representing the ML model. This circuit is then used to generate ZKPs for user inferences when needed. The ZKPs can be sent to the user for verification or posted to a public blockchain that handles the verification task. If the ML model is private, independent third parties can verify that the utilized zero-knowledge circuit accurately represents the model.
This trust minimization aspect of ML models is particularly useful in high-stakes scenarios. For example:
In medical diagnoses, a patient could submit their data to an ML model for potential diagnosis. They need guarantees that the target ML model was correctly applied to their data. The inference process generates a ZKP that proves the correct execution of the ML model.
In assessing creditworthiness for loans, ZKPs can ensure that all financial information submitted by an applicant is considered by banks and financial institutions. They can also demonstrate fairness by proving that the same model is used for all users.
In insurance claim processing, which is currently manual and subjective, ML models could better assess claims considering the policy and claim details. Combined with ZKPs, these claim-processing ML models can be proven to have considered all policy and claim details and that the same model is used to process all claims under the same insurance policy.
* There are more examples illustrated here as I can see this area to be one of the first areas where we could see meaningful participation of crypto in the AI revolution.
AI Services Marketplace
In a world where AI capabilities are increasingly in demand, a decentralized AI marketplace could prove to be the next big thing. Developers, researchers, and organizations could offer their specialized AI models or services for others to use or purchase.
For instance, consider a scenario where a collective has trained an AI model that can accurately predict weather patterns, crop harvesting timelines and demand forecasting based on a variety of data inputs. Traditionally, this model might only be accessible to a specific organization or limited to a single application. However, in a decentralized AI marketplace, this model could be shared with or sold to anyone interested in Agritech. This could be a farmer looking to optimize their planting schedule, a logistics company planning their routes, or retail players looking to optimise sales strategies.
The integration of crypto could enable microtransactions for AI services, particularly those with low transaction fees. For example, users could pay a small fee in cryptocurrency to use an AI model for a single task, instead of needing a subscription or large upfront cost. This could make AI services more accessible to a wider range of users.
Another potential application is AI-driven financial services. AI could provide personalized financial advice or investment strategies based on an individual's financial situation and risk tolerance. Users could pay for these services in crypto, and potentially even execute the advised transactions (such as buying or selling assets) directly on a blockchain platform.
The decentralized marketplace not only creates a new revenue stream for the developers but also democratizes access to high-quality AI models, allowing smaller entities and individuals to leverage AI capabilities that would otherwise require substantial resources to develop – further optimising value in the overall supply chain. It can also lead to an increased diversity of AI models available. Each developer or organization may have their own unique approach to training AI models, and this variety can foster innovation and prevent a monopoly of ideas.
Consider another scenario where an organization has developed a powerful AI for disease diagnosis based on medical imaging. In a decentralized AI marketplace, this tool could be made available to hospitals and clinics around the world, potentially saving lives in regions where such technology was previously inaccessible due to resource constraints or geographic limitations.
Moreover, a decentralized AI marketplace could facilitate collaboration and collective problem-solving. If an AI model needs improvement or doesn't quite fit the requirements of a user, multiple parties could potentially collaborate to refine and adapt the model.
IP Ownership
Crypto can help address some of the challenges associated with data privacy, ownership, and trust in the rapidly evolving world of AI. Users can maintain control over their data and decide which AI services can access it, potentially receiving crypto payments in exchange.
One notable area where this intersection becomes particularly significant is in the realm of AI-generated content. As this becomes more prevalent, verifying the authenticity of such content becomes critical to avoid the advancements of deepfake content to manipulate digital media. Here, cryptographic methods such as Zero-Knowledge Proofs (ZKPs), can play a key role. They can be used to verify the outputs of AI services without needing to trust the service provider, adding a layer of trustless verification to the system. This could be particularly useful in fields like machine learning model verification, where ensuring the accuracy and reliability of models is crucial.
Moreover, AI has the capability to generate a wide variety of content, from articles and music to artwork. This content could be sold or licensed in a decentralized marketplace, with transactions facilitated by crypto. This approach not only opens up new avenues for creators to monetize their AI models, but also allows users to access unique AI-generated content.
Crypto networks can facilitate fair compensation of royalties to content creators, ensuring that creators receive their deserved share of revenue. This not only promotes a more equitable distribution of rewards, but also incentivizes the continued creation and development of high-quality AI models and services.
AI Governance through DAOs
Decentralized Autonomous Organizations (DAOs) offer an innovative approach to AI governance. DAOs can harness AI to streamline and democratize decision-making processes, as token holders vote on the parameters governing the AI's operations. This structure enhances efficiency, enabling the digital economy to scale effectively.
DAOs could also contribute to AI in multiple ways, fostering a collaborative environment for AI development, training, and execution. For instance, DAOs could facilitate the pooling of resources, allowing members to collectively train and fine-tune AI models. They could also form a consensus on the ethical guidelines that AI should follow, helping to shape AI systems that align with societal values and norms. Moreover, DAOs could facilitate the sharing and monetization of AI services in a decentralized marketplace, promoting innovation and broadening access to AI capabilities.
In essence, DAOs serve as the organizing structures within the digital business landscape, guiding the growth and application of AI. Their robust governance frameworks help ensure that AI technologies are managed responsibly and effectively, maximizing benefits and minimizing potential harm. Envisioning the digital economy, one might draw an analogy that DAOs are akin to business entities, AI represents the businesses themselves, and cryptocurrencies serve as the national reserve currency. Just as business entities structure the commercial activities, DAOs provide the organized framework for AI technologies to operate with crypto being the national reserve currency that underpin the value exchange within this ecosystem, enabling transactions and value storage.
Conclusion
The convergence of AI and crypto signifies a significant shift in our technological landscape, potentially transforming various aspects of our digital economy. The decentralized nature of cryptocurrencies aligns with the shared, inclusive vision of AI, promoting democratization and access to advanced services.
Crypto can provide a means to transact in this new economy, enabling microtransactions and fostering a marketplace for AI services and content. Meanwhile, blockchain technologies like zero-knowledge proofs and DAOs bring trust, transparency, and structure, ensuring that AI deployment is fair, verifiable, and organized.
However, this convergence isn't a simple fusion. It's a fundamental change in how we interact with technology, reshaping our concepts of business and economics in the digital age. With thoughtful regulation and ethical considerations, this convergence could help pave the way to a more efficient and equitable digital future. As we continue to explore the potential of AI and crypto, it's crucial that we navigate this new terrain with caution, inclusivity, and a keen eye on the societal implications.
Thank you for reading our thoughts on AI and the crypto economy. If you enjoyed this article, consider giving us a follow on our Twitter, as well as the author of this piece, Za’im Zainudin.