The privacy computing platform Oasis Network has been revitalized: Smart privacy born for Web3 and AI

Author: Oasis Network

We are in the midst of a revolution with generative artificial intelligence (GenAI) and large language models (LLMs), which are rapidly transforming every aspect of our lives. The current demand is for technology to enable the decentralized, trustless, and democratized development and deployment of AI systems.

Since the launch of the Oasis mainnet in 2018, Oasis's vision has been to build a privacy-first blockchain network to drive a decentralized data economy. Oasis has never strayed from our steadfast focus on the use of data and derived data products (such as AI models), which embodies and enhances the spirit of transparency, agency, and decentralized governance in Web3.

Under the leadership of the Oasis community, we have launched a brand refresh campaign and provided more support for AI dApps that offer programmable privacy for users. We are also honored that one of the top AI-driven protocols in blockchain, Ocean Protocol, has chosen us as their preferred privacy infrastructure for developing Ocean Predictoor, a prediction market natively built on Oasis Sapphire. Since its launch in September 2023, Predictoor has already reached nearly $1 billion in monthly trading volume.

In this article, we will take you into the decentralized AI we envision.

Building a More Decentralized Digital Future

New technological developments allow mature blockchain platforms to extend trustless computing to real-world infrastructure, traditional data assets, and the AI domain. However, in the context of Web3, these advancements will not succeed if they do not maintain the principles of Web3 through sufficient decentralization.

Oasis can ensure the verifiability and privacy of arbitrary computations in a decentralized environment. From autonomous AI agents, NFTs with their own thoughts, decentralized AI training, oracles, to chain abstraction and deep defense, privacy-driven AI has endless possibilities.

On the engineering front, Oasis helps AI dApp teams add privacy features for their users by adding additional functionalities on the verified Oasis Sapphire stack.

ROFL (Runtime Off-Chain Logic): Decentralized Privacy Built for AI

As generative AI and AI-as-a-service become ubiquitous, there are risks such as privilege escalation and cross-tenant attacks on inference pipelines. Protecting these AI pipelines using privacy computing technology is crucial for model training, whether using open data sources or licensed data sources, as well as inference pipelines. Privacy computing using Trusted Execution Environments (TEEs) allows training and inference algorithms to run in encrypted memory with runtime integrity, making their functionality less susceptible to tampering by any other external cloud services.

Runtime Off-Chain Logic (ROFL) is a framework that adds privacy computing support for off-chain components to runtimes like Oasis Sapphire, enabling non-deterministic behaviors such as using random algorithms and accessing network resources while providing the privacy and runtime integrity required by AI pipelines. ROFL allows off-chain components to communicate seamlessly with off-chain domains, achieving full composability between different blockchain platforms and offline computing stacks. This not only tracks the sources of data used for AI training but also tracks the origins of AI models used as derived data products in inference pipelines, while providing privacy and integrity for data providers, model providers, and even AI-as-a-service providers.

These components are protected by the same Oasis Trusted Execution Environment (TEE), consensus layer, and its decentralized validator set, which can transparently run ROFL without the need to install anything other than the Oasis core nodes and runtime bundles (required for current node operation). ROFL can be added to any privacy runtime to extend its capabilities, whether existing or new runtimes.

To learn more about how Oasis deploying ROFL opens new possibilities for AI, such as how AI-driven decentralized agents can have private "thoughts," please refer to the blog written by Oasis Director Jernej Kos.

Collaboration Between Oasis and AI Teams

Ocean Predictoor ------ Trading Prediction dApp

Ocean Predictoor is a crowdsourced, on-chain dApp for making and validating predictions. Individuals can submit predictions and stake on them, earning funds when their predictions are correct, and losing funds otherwise. Traders can purchase aggregated accurate predictions in the Predictoor dApp and use them to buy or sell certain assets or contracts.

Leveraging the privacy of the Sapphire network, Ocean relies on Oasis technology to ensure the execution of its AI-driven on-chain marketplace. Without this privacy, the information collected and shared through Ocean oracles and AI products would be exposed to everyone. The success of the new Ocean Predictoor dApp has established a strong partnership between the two teams. Now, we are further deepening our collaboration with Ocean to expand our achievements.

Currently, Ocean has partnered with SingularityNET and FetchAI to establish the Artificial Superintelligence Alliance (ASI) to continue open-source AI research and development. They are committed to building a decentralized AI infrastructure as a viable alternative to the centralized AI ecosystem dominated by big tech companies.

deltaDAO ------ AI Data Management

Recognizing the advantages of the Oasis runtime—separating consensus and computation layers to replicate a privacy computing environment with shared state—deltaDAO has launched its Pontus-X ecosystem, transforming the way traditional enterprises achieve AI and data product tokenization, and helping them transition from Web2 to Web3, where they can fully leverage digital goods while retaining control over their data.

The launch of Pontus-X improves the way different industries share and tokenize their data assets. It breaks down the data silos that exist in the Web2 world, unleashing the power of AI and data. Pontus-X aims to simplify the way companies tokenize their data while retaining complete control and privacy, leveraging the unique capabilities of our network. Through end-to-end data exchange and service revenue, deltaDAO is building a data ecosystem to ensure regulatory clarity for businesses and users when training and using AI, providing trusted solutions.

Why Privacy is Crucial for AI

AI requires data sources for training, and while some data is publicly available and usable, much of it is the intellectual property of data providers or individuals. The use of this data may require specific permissions or involve compensation. Additionally, it involves a significant amount of computation, which blockchain technology alone cannot provide. Oasis uniquely meets these needs through its privacy features and framework.

The Oasis team is developing a new product that draws on the previous successes of the consensus layer and Sapphire ParaTime, combining them with verifiable off-chain computing, which utilizes the same foundational cryptography of today's Oasis network. Verifiable off-chain capabilities will address the demands of compute-intensive workloads in AI, as well as the data confidentiality and runtime integrity required when using proprietary data for AI training. We believe this is the cornerstone of decentralized AI.

The Future of Runtime

Trusted Execution Environment (TEE) runtime extensions can perform complex computations in off-chain privacy environments while allowing on-chain verification of execution integrity. This paves the way for truly decentralized AI models with verifiability and privacy. For example, AI models trained for facial recognition can securely handle sensitive data, while Sapphire's smart privacy features ensure that data contributors are compensated for their data and that AI models are unbiased. Furthermore, this can also allow for verification of compliance with data protection laws, which is particularly important in industries like healthcare, where the privacy and security of patient information must be protected. Oasis's decentralized privacy computing platform ensures that AI models run securely to prevent unauthorized access to sensitive information.

Support for Data Management

Oasis's decentralized key management system ensures controlled and auditable data access. This feature enables secure data collaboration across industries, including finance, where encrypted data can be analyzed to extract valuable insights without exposing underlying information. The secure environment and decentralized key management allow for encrypted verification of the source and integrity of the provided data, combating deepfakes and unauthorized content replication.

This is just the beginning of Oasis's journey in the AI space, and we have limitless possibilities ahead. If you wish to build on Oasis and learn how decentralized AI and smart privacy can reach their full potential, as well as stay updated on relevant hackathon activities, please visit our official website.