Exponential Technology Stack Convergence: Generative AI-enabled Blockchain Networks

Throughout 2024, we expect disproportionate disruption when the exponential tech stack starts to converge. One of the most promising of these exponential technological intersections will be Generative AI-enabled Blockchain Networks – of which applications and use cases have started to emerge.

Background

OODA Almanac 2024: The Exponential Tech Stack Starts to Converge

Regular readers of the OODA Loop know that we cover exponential technologies daily and we expect disproportionate disruption when these technologies start to converge. For example, AI + Bio-Tech or Robotics + AI. We will be tracking and continue re-orienting you to developments in the following areas:

Quantum Tech: This is the ultimate in first principles engineering. With new insights into how the quantum world works this is becoming a foundational science for all other engineering disciplines. Quantum Computing may be a decade away, but quantum engineering is a reality today resulting in more powerful microelectronics, more capable sensors, and improved cybersecurity solutions.
Bio-Tech: Until this day, all biological science was based on observation and experimentation. New BioTech enables the application of engineering principles to life itself. In 2024 we expect Bio-Tech to continue to improve health and pharmaceutical outcomes and to start disrupting fields such as mining, manufacturing, agriculture, and energy. Watch for mainstreaming of Brain Machine Interfaces towards the end of the year.
Narrow AI: The next year will bring more sophisticated narrow AI applications like OpenAI’s ChatGPT into areas like healthcare diagnostics, marketing, and customer service. Employee disruption is already well underway. Companies, governments, and individuals will adopt or not (“Adopt or you’re toast”).
General AI: General AI is a term used to describe technology so sophisticated that it can solve things across multiple domains, like a human. We do not believe reaching a General AI is a simple binary event. We will more likely see a continued improvement in multiple AI tools in 2024. Prepare to be amazed.
Advanced Robotics and Automation: The most advanced robots are giving physical form to AI. In 2024 we expect to see humanoid robots in manufacturing and warehousing. In 2025 some of your neighbors will have them in their homes.. Autonomous vehicles and drones are posed to disrupt transportation and logistics.
Materials Science: Innovations in materials science, particularly in additive manufacturing and 3D printing, will lead to more sustainable and efficient manufacturing processes across multiple industries in 2024. The cost of capital to modernize industries is inflationary, but the ability to manufacture in new ways with automation is a long-term deflationary trend.
AR, VR, and the Metaverse: Augmented and virtual reality technologies are becoming more immersive, making the metaverse a more integral part of entertainment, education, and remote work. The Apple Vision Pro is the latest in a long evolution of these technologies.
Space Technologies: In the coming year we will witness new milestones in space technology, opening new avenues for pharmaceutical production, earth observation, telecommunications, and human space travel.
Blockchain and Distributed Ledger Technologies: OODA has been tracking this domain closely and sees the foundations being laid for new applications across multiple aspects of society. Solutions will accelerate in domains like finance, healthcare, security, supply chain management and even voting systems. One measure of potential disruption in this domain is the number of developers creating blockchain-based solutions. There were 22,000 blockchain developers in the US in 2022. By the end of 2024, we expect that number to more than double.

Generative AI-enabled Blockchain Networks: Fundamentals, Applications, and Case Study

You have to sort the wheat from the chaff regarding the infinite amount of academic research papers available on blockchain and generative AI. The cream usually rises to the top – as is the case with this research by this global team predominantly from Southeast Asia:

Cong T. Nguyen is with the Institute of Fundamental and Applied Sciences, Duy Tan University, Vietnam.
Liu, H. Du, and D. Niyato are with the School of Computer Science and Engineering, Nanyang Technological University, Singapore
Dinh Thai Hoang and Diep N. Nguyen are with the School of Electrical and Data Engineering, University of Technology Sydney, Australia
S. Mao is with the Department of Electrical and Computer Engineering, Auburn University, Auburn, USA.

This research paper represents the role academia and network-based global collaboration play in new market creation – i.e. the creation of a new technology marketplace and ecosystem – complete with “early adopter” ideas for business model generation, value creation design, and early use cases.

If we revert to first principles as the foundation of the future; the first principle of the design and launch of DARPAnet – what is now the internet – was to enable scientists from around the world to more easily collaborate and share research findings and scientific datasets at the speed of light. This first principle is proving to be one of the only quantifiable acts of constant global collective intelligence that can scale into a scientific response to a crisis based on the severity of the challenge if needed (see the genomic data shared in the run-up to the creation of the mRNA Covid vaccine).

But it is the real-time cross-pollination of global ideas and the instantaneous availability of research from across the globe that we track as one of the only structural efforts that may well be the saving grace for finding solutions or convergences of technology to apply in this era of global polycrisis.

Simply put: The infrastructure is in place for a global, science and solutions-based approach to the more dystopian unintended consequences of exponential technologies or the climate crisis – to name just a couple of the current crises. We continue to track if this global, internet-based response to these challenges will be adequate to match the exponential speed of emerging technologies and the complexities of physics and climate science.

For now, we return to global research’s role in the formation of new technology markets – specifically, potential applications and use cases in the convergence of AI and Blockchain.

From the white paper:

Abstract

“Generative Artificial Intelligence (GAI) has recently emerged as a promising solution to address critical challenges of blockchain technology, including scalability, security, privacy, and interoperability. This paper introduces GAI techniques, outlines their applications, and discusses existing solutions for integrating GAI into blockchains. Then, we discuss emerging solutions that demonstrate GAI’s effectiveness in addressing various blockchain challenges, such as detecting unknown blockchain attacks and smart contract vulnerabilities, designing key secret sharing schemes, and enhancing privacy. Moreover, we present a case study to demonstrate that GAI, specifically the generative diffusion model, can be employed to optimize blockchain network performance metrics.

Experimental results clearly show that, compared to a baseline traditional AI approach, the proposed generative diffusion model approach can converge faster, achieve higher rewards, and significantly improve the throughput and latency of the blockchain network.” Additionally, we highlight future research directions for GAI in blockchain applications, including personalized GAI-enabled blockchains, GAI-blockchain synergy, and privacy and security considerations within blockchain ecosystems.”

Introduction

“Blockchain technology, renowned for its exceptional ability to maintain data integrity and immutability in decentralized settings, has been increasingly recognized as a crucial enabler for transparent data management. Fundamentally, a blockchain operates as a distributed ledger, where records are collectively maintained and shared across a peer-to-peer network. This technology leverages sophisticated cryptographic methods and consensus mechanisms to provide self-governance, security, transparency, and efficiency. Its applications are diverse and far-reaching, encompassing sectors such as finance and healthcare, and extending to innovative domains like the Metaverse and Web 3.0…While blockchain technology is innovative, it faces challenges such as scalability, security, privacy, and interoperability…the integration of traditional Discriminative Artificial Intelligence (DAI) with blockchain shows great promise in addressing these issues…

…as AI technologies continue to evolve, Generative AI (GAI) has recently been emerging as a focal point, drawing even greater levels of attention…compared to DAI, GAI offers a distinct advantage thanks to its outstanding ability to generate data, coupled with its innate creativity and flexibility. Particularly, GAI can create realistic content by using latent vectors to represent given samples and learning their distribution. This allows GAI to overcome data scarcity by synthesizing new data…”

Discriminative AI (DAI) and Generative AI (GAI)

The researchers highlight the key distinctions between Generative AI (GAI) and Discriminative AI (DAI). GAI focuses on modeling data distribution and can enhance performance by generating more content. On the other hand, DAI concentrates on modeling the relationship between data and its labels, such as classifying images based on characteristics. These distinctions show how GAI aims to replicate data patterns to generate new content, while DAI focuses on the association between data and labels like classification tasks. GAI’s approach lies in understanding data distribution, enabling it to progressively improve generated content quality, while DAI emphasizes the data-label relationship for tasks like image classification. This distinction showcases how GAI and DAI differ in their fundamental approaches to handling and utilizing data within AI systems.

The integration of Discriminative Artificial Intelligence (DAI) with blockchain technology. DAI can enhance scalability, refine consensus mechanisms, optimize network resource allocation, monitor transaction history for fraud detection, and be collectively trained by users for platform security. DAI can be utilized to identify vulnerabilities in smart contracts and known risky behavior in contract code. However, when labeled data is limited, DAI techniques may face challenges in their effectiveness. The combination of DAI and blockchain leverages the strengths of each technology to enhance overall integrity and security.

On the other hand, Generative Artificial Intelligence (GAI) can be used to generate adversarial inputs to test smart contracts under various conditions, aiding in the detection of unknown vulnerabilities that may surpass the capabilities of DAI. Generative AI (GAI) is highlighted in the report as a promising solution for addressing challenges in blockchain technology, including scalability, security, privacy, and interoperability. GAI is praised for its creativity, flexibility, and ability to generate new content based on training and user inputs. This technology can analyze patterns and structures in existing data, making it effective in tackling various blockchain challenges that traditional AI approaches may struggle with. Compared to traditional AI, GAI stands out for its capacity to create new information autonomously, such as images, texts, sounds, videos, and system designs, by learning from existing data. Its strengths lie in generating data, creativity, and adaptability to novel scenarios, making it a valuable asset in addressing emerging needs in blockchain development.

GAI Models with High Potential for Blockchain Networks

Variational Autoencoder (VAE) is highlighted for its unique architecture that enables it to learn compact data representations within a lower-dimensional latent space. This efficiency makes VAE highly suitable for generating data based on long-term distributions, such as transaction history or smart contract usage. Additionally, VAE’s encoder network maps input data into a latent space distribution, while the decoder network generates data samples from this distribution, showcasing its potential to contribute significantly to blockchain networks.

Generative Adversarial Networks (GANs) show high potential for blockchain networks due to their ability to generate high-quality data essential for training blockchain attack detection systems. GANs excel at creating synthetic data samples and distinguishing between real and fake data through adversarial training, making them valuable for simulation and evaluation in the blockchain. Additionally, GANs, along with other Generative AI techniques, can address challenges in blockchain networks such as scalability, security, privacy, and interoperability. Leveraging GANs can optimize blockchain network performance by improving throughput and latency compared to traditional AI approaches. Through the integration of GANs and Generative AI into blockchain networks, there is a potential for enhanced data generation, system performance, and future research advancements.

Generative Diffusion Model (GDM) offers high potential for blockchain networks by optimizing performance metrics like throughput and latency. GDM iteratively adds noise to data points and denoises them to converge to the desired data distribution, making it suitable for generating high-quality data for blockchain optimization GDM’s ability to produce realistic and diverse data samples contributes to its application in generating solutions for blockchain optimization problems.

Large Language Models (LLMs) have high potential for blockchain networks due to their ability to understand and generate text-based data, such as smart contract code. LLMs are AI models built on deep neural networks with millions to billions of parameters, enabling them to capture and generate human-like text across various languages and topics. LLMs can help in improving blockchain networks by generating high-quality data for evaluation and providing solutions to blockchain optimization problems. Additionally, LLMs can produce diverse and novel content related to the training data, which can lead to faster convergence, higher rewards, and improved throughput and latency in the blockchain network. The integration of LLMs with blockchain can enhance scalability, optimize network resource allocation, and refine consensus mechanism design.

From the report – TABLE I: Comparisons of the applications of DAI and GAI in blockchain networks

From the report – TABLE II: Summary of GAI approaches for blockchain.

Case Study: Diffusion Model-based Blockchain Design

By leveraging a Generative Diffusion Model (GDM) to optimize blockchain systems, the study demonstrates through a case study how GAI can assist in enhancing a blockchain’s performance by optimizing metrics such as throughput and latency. By leveraging GDM, blockchain systems can be optimized with faster convergence, higher rewards, and improved network performance compared to traditional AI approaches. The research further explores the potential applications and solutions for integrating GAI into blockchain networks, addressing challenges like scalability, security, privacy, and interoperability.

What Next?

The synergy between Generative Artificial Intelligence (GAI) and blockchain technology offers promising solutions to address challenges faced by blockchain systems:

By leveraging GAI techniques like the Generative Diffusion Model (GDM), blockchain network performance can be optimized to achieve faster convergence, higher rewards, and improved throughput and latency; and
The collaboration between GAI and blockchain can establish a continuous cycle of improvement, enhancing the privacy, security, and trustworthiness of GAI models and training processes.

Future research directions include:

Personalized GAI-enabled blockchains
GAI-blockchain synergy, and
Privacy and security considerations within blockchain ecosystems.

For the full white paper, go to Generative AI-enabled Blockchain Networks: Fundamentals, Applications, and Case Study

AI + Blockchains Come Together

Marc Andreesen’s shop, 16z, in collaboration with Stanford Professor Andrew Hall, included the convergence of AI and Blockchain in their list of “Big Ideas in Tech for 2024”:

“Decentralized blockchains are a counterbalancing force to centralized AI. AI models (like in ChatGPT) can currently only be trained and operated by a handful of tech giants, since the required compute and training data are prohibitive for smaller players. But with crypto, it becomes possible to create multi-sided, global, permissionless markets where anyone can contribute — and be compensated — for contributing compute or a new dataset to the network for someone who needs it. Tapping into this long tail of resources will allow these markets to drive down the costs of AI, making it more accessible.

But as AI revolutionizes the way we produce information — changing society, culture, politics, and the economy — it also creates a world of abundant AI-generated content, including deep fakes. Crypto technology can be used here as well to open the black box; track the origin of things we see online; and much more. We also need to figure out ways to decentralize generative AI and govern it democratically, so that no one actor ends up with the power to decide for all others; web3 is the laboratory for figuring out how. Decentralized, open-source crypto networks will democratize (vs. concentrate) AI innovation, ultimately making it safer for consumers.”