Becoming the Backbone

Gateway is poised to revolutionize the decentralized AI industry by becoming its "Layer 0"—the foundational backbone that ensures data security, privacy, and user sovereignty from the ground up. As the world moves towards increasingly sophisticated AI technologies, the need for a robust, decentralized infrastructure that can handle massive volumes of data securely and efficiently has never been more critical.

At the heart of Gateway's innovation is its capability to provide secure, private-by-default decentralized storage via Encrypted Data Vaults (EDVs). Gateway’s architecture embeds native access control, empowering users with unparalleled control over their data. With Gateway, data sovereignty is not just a feature; it is a fundamental principle. Users can directly manage who accesses their data and under what conditions, using smart contracts and advanced permission settings that are verifiable and immutable. This capability is critical for industries where data sensitivity is paramount, such as healthcare and finance, providing a secure environment for sharing and analyzing data without exposing it to risks.

Gateway's role as the "Layer 0" for decentralized AI extends beyond just secure storage and privacy. It facilitates a scalable and interoperable environment where AI applications can operate across different platforms and ecosystems without the typical constraints associated with centralized systems. Gateway's protocol is designed to handle the high demands of AI workloads, from training machine learning models on large datasets to performing real-time AI inference at scale.

Building with Gateway: A Necessity for Decentralized AI/Inference/LLM Networks

In the wake of high-profile data breaches and the increasing scrutiny of data privacy practices, the demand for robust data management solutions has become more pressing than ever. Centralized data handling poses significant risks, as evidenced by incidents such as the recent AT&T T-Mobile leaks this year, the Facebook-Cambridge Analytica scandal, Meta-Netflix privacy collusion, Bank of America data breaches, and 23andMe’s ancestry data hack. This exposes the vulnerabilities and potential misuse of personal information in centralized systems. While emphasizing why decentralized AI and LLM networks seek architectures that not only enhance security and privacy but also maintain high performance and scalability.

Gateway addresses these needs by providing an infrastructure that is inherently secure, fully interoperable, and capable of supporting large-scale AI applications. Here are the core principles of why build with Gateway.

1. Enabling Users to Become Databases with EDVs

Many decentralized networks are focusing on avoiding major risks with handling highly sensitive and proprietary data in centralized databases… instead, it is time to enable a web where each user becomes their own database. Gateway upholds the principle of privacy by design. This is particularly beneficial for AI networks that handle sensitive or regulatory-bound data, assuring that data privacy is maintained without compromising on accessibility for legitimate processing needs.

In their Litepaper @gensynai mentioned “Whilst large amounts of ML research can be performed on open datasets, final model fine-tuning often uses proprietary user data. More specifically, in our interviews with ML engineers and CTOs, they indicated that data privacy was orders of magnitude more important than model privacy.” Gateway facilitates encrypted, secure, and private data to train the next generation of tailored ML models.

Gateway’s decentralized storage model shards data across multiple EDVs. This not only enhances security by mitigating risks associated with central data repositories but also improves data availability and fault tolerance. In the event of a node failure, the system can retrieve and reconstruct data from other nodes without any data loss due to erasure coding, ensuring continuous availability for AI computations and model training.

2. Your Keys, Your Choice: Enabling Access Control

The principle of user sovereignty is fundamental in decentralized networks, where users should have control over their data. This is true, especially in networks that hope to align incentive and growth mechanisms between their network usage and user participation. Gateway's smart contract-based access control mechanisms empower users to define and enforce how their data is accessed and used, enabling private data to empower intelligent machines. This reinforces user trust and aligns with stringent global data protection regulations such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA). By integrating Gateway, decentralized AI networks can provide their users with transparent, verifiable, and enforceable data management tools, thereby enhancing user engagement and trust.

3. Gateway Chain: The Global Data Trail

While blockchains are mostly used for financial applications, the use cases in portable data are similarly crucial. Gateway’s L1 blockchain ensures that all data transactions and accesses are transparently recorded and immutable. This audibility is crucial for compliance with regulatory standards and for maintaining an open and trustworthy relationship with users and regulators alike. Gateway’s chain tracks transactions such as identity creation, data contributions, data shares, access revocation, and more.

Scalability is a critical concern for AI and LLM platforms, especially those requiring real-time data processing. Traditional blockchain solutions often face challenges in scaling efficiently. Gateway's protocol design addresses these challenges by offering a scalable architecture that supports high-throughput and low-latency operations essential for AI workloads. Additionally, Gateway's commitment to interoperability allows AI networks to operate seamlessly across different blockchain ecosystems, facilitating broader application and integration possibilities.

With Gateway, the focus for AI teams is two-pronged. The first is contributing critical data such as generation outputs, inference models, and proprietary copyright information. The second is being able to quickly request and retrieve personalized data to train algorithms across various networks while not being burdened with exorbitant fees and high latency.

4. A Future of Encrypted Computation

Gateway is thinking of a future where data is an asset. An increasingly more valuable one, as the AI industry is fully dependent upon the information available to train models. However, in the same way that you do not share access to your bank account to get loans, we do not want to necessitate sharing access to your data vault to join in on the growth of AI. How?

Gateway is exploring advanced cryptographic techniques such as Zero-Knowledge Proofs (zkProofs) and Fully Homomorphic Encryption (FHE) to ensure that data remains encrypted and secure throughout its lifecycle, including during computation processes. This approach allows AI and LLM networks to perform complex computations on data without ever exposing the underlying data, thus safeguarding user privacy and enhancing security against potential breaches.

This is why Gateway is exploring computational capabilities with teams like Zama. In a conversation with Zama, CEO Rand stated "What Gateway is doing is great. So I'm really pushing for someone to build what they’re building for private data and distribution. It has to exist. And I think FHE makes so much sense in it because you cannot have all the data public.”

Process Flow for an AI Inference Network Using Gateway

In this example let us dive into the convergence between the Traditional Credit Score Company (CreditFi), Decentralized AI Agents training financial models (ML Finance), and a data owner (Jerry) who wants to share his data, get quality products, and join in on his favorite Crypto AI project’s growth.

Step 1: Make Data Available to Jerry

1. Data Contribution + Encryption: CreditFi contributes data to a user’s data vault. Prior to uploading to the Gateway network, data is encrypted using public keys associated to the data owner, Jerry. This ensures that data is protected and only accessible by entities with the appropriate permissions or fulfilled payment obligations.

Step 2: Data Storage Using Encrypted Data Vaults (EDVs)

1. Storing Encrypted Data: The encrypted data is stored in EDVs within the Gateway network. These vaults are equipped with strong security measures to protect data against unauthorized access.
2. Data Assurance Commitments (DACs): These commitments ensure the integrity and availability of the stored data, managing resource allocation appropriately on the network.

Step 3: Jerry Trains ML Finance

1. Access Control: Using Gateway’s advanced cryptographic features like proxy re-encryption, Jerry can grant conditional access to his data for specific purposes, such as training an ML Finances advanced personal financial advisor.
2. Decentralized Model Training: AI models can be deployed directly on top of the Gateway network through external but connected nodes that interact securely with stored data. Jerry manages requests under strict cryptographic controls ensuring data privacy and security.

Step 4: ML Finance Requests Data

1. Data Processing: When AI inference is required, ML Finance requests access to specific datasets — in this case, Jerry’s investment history and borrowing history. This access can be granted via smart contracts that enforce the terms of data usage.
2. Secure FHE Computation: For further privacy, Gateway could facilitate secure Fully Homomorphic Encrypted computation techniques, allowing multiple parties to compute results based on shared data without exposing the underlying data itself.

Step 5: ML Finance Helps Jerry

1. Generating Results: After processing the data shared. Jerry receives the financial advice he was looking for based on the personalized information he has shared.
2. Aligning Incentives: ML Finance wants Jerry to continually share his financial data and generate inputs into their network, so they actively request his updated financial information and reward him with native tokens/rewards as their Decentralized Network grows
3. Storing Inferences: The results (e.g., AI-generated insights) are encrypted and written to Jerry’s EDV, allowing him to control how other applications or networks may build open ML Finance’s outputs.

Key Points: User Autonomy and Data Management

1. User Autonomy: Users retain control over their data through features like key revocation and data deletion capabilities within the Gateway protocol. They can revoke access at any time, ensuring compliance with data protection regulations (e.g., GDPR).
2. Controlled Data Sharing: Results or even insights derived from the data can be shared across the web using Gateway’s secure sharing mechanisms, ensuring that data privacy is maintained throughout its lifecycle.
3. Audit and Compliance: All transactions, access logs, and data usage are recorded on the blockchain, providing a transparent and immutable audit trail that enhances trust and compliance.

Forging the Gateway to AI Today…

Gateway is excited that we are building with leading teams in the Decentralized/Crypto AI space. Here is a sneak peek into a few AI teams building with Gateway today!

1. Atoma Network

Atoma is building the next-generation decentralized AI inference network. Through a novel consensus design, Atoma will empower Web3 applications, as well as governments and enterprises, with its layer of verifiable AI inference. Atoma is creating a world where AI will be our co-pilot, caretaker and companion, trust in generative AI infrastructure will be of critical importance. Through Atoma, a new paradigm of governance, trust and democratization to the AI community, will enable the development of open, inclusive and permissionless AI applications.

Hashim Khan, co-founder of Atoma Network, stated “I'm most excited about Gateway because it aligns perfectly with our vision at Atoma. The ability to securely manage data while maintaining user control not only enhances our Inference Network but also builds trust and functionality for our network. Soon users will be able to share inference data with our applications and beyond!”

2. Heurist

Heurist is a layer 2 network for AI model hosting and inference. Think of them as the HuggingFace of Web3. Heurist aggregates compute resources from individual GPU owners and allow ecosystem developers to build AI apps with API access to open source AI models.

JW co-founder of Heurist stated “At Heurist we are big believers in Gateway's "each user is their own database" ethos. With Gateway's privacy-first and infrastructure, We can give back ownership to users who express their creativity with Heurist's open-source image-generation models. Creators can monetize their creations, manage copyright, and even showcase a metaverse gallery in an on-chain provable way".

In conclusion, Gateway is not merely a platform but a paradigm shift in how we approach data in the AI-driven future. By prioritizing security, privacy, and user sovereignty, Gateway is set to become the indispensable "Layer 0" for the decentralized AI industry, enabling smarter, safer, and more efficient AI applications that respect user data and foster innovation. Becoming the bridge that allows users to participate and control their experiences in the fastest-growing industry in the world. As we stand on the brink of this technological revolution, Gateway offers the promise of a decentralized, user-centric world where AI can thrive in harmony with data privacy and security.

What are you waiting for anon, get started with Gateway right now

https://docs.mygateway.xyz/developer-guide/sdk/introduction

Author:
Ayyan Rahman
Co-Founder @Gateway