Project Overview

Sahara AI is a decentralized blockchain platform focused on making AI technology more accessible to individuals and businesses. It aims to democratize artificial intelligence using decentralized methods. The platform emphasizes fairness, transparency, and privacy protection, offering decentralized ownership and governance mechanisms to ensure that contributors receive the recognition and rewards they deserve.

Decoding Decentralized AI: Sahara AI

Project Roadmap

On August 26, 2024, Sahara AI, the AI blockchain platform, unveiled its roadmap. The first phase of the Sahara AI Marketplace will begin in the third quarter of 2024 to facilitate collaboration on knowledge bases. In the fourth quarter, a testnet for Sahara Chain will be launched, allowing developers and users to test and validate its features. The mainnet for Sahara Chain is set to launch in the first quarter of 2025, enabling the operation of decentralized AI applications. By the second quarter of 2025, Sahara agents will be deployed, equipped to perform complex tasks and make decisions using decentralized data through advanced AI capabilities.

Source: Official X account of Sahara AI

Project Background

The Sahara AI project is rooted in the issue of power concentration in the artificial intelligence field. Currently, a few major tech companies dominate the development, deployment, and monetization of AI models and applications, hoarding knowledge and creativity from users, developers, and innovators. Unfortunately, these contributors often receive little to no recognition or compensation for their efforts. This inequitable system stifles smaller innovators, whose creativity inadvertently fuels the growth of large corporations without any benefit to them. As a response, Sahara AI aims to create a fairer and more transparent AI ecosystem, ensuring that all participants can share in the value and benefits generated by AI development. The project seeks to prevent privacy violations and excessive power concentration while promoting the broader societal use of AI’s potential.

Project Advantages

Most AI technologies today remain entrenched in the Web 2.0 model, where users trade their knowledge contributions for access to AI and machine learning tools. Centralized AI providers often exploit users’ proprietary models and data in opaque ways, failing to protect privacy or offer fair compensation. Sahara AI presents a human-centered alternative that fundamentally changes this model by prioritizing personal privacy and rewarding all contributors. The main advantages of Sahara AI can be summarized in three key areas:

• Ownership and Attribution: Sahara AI ensures that ownership and governance of artificial intelligence are decentralized throughout the entire development process, with clear and immutable attribution recorded on the blockchain.
• AI Utility Tools: Sahara AI offers a comprehensive suite of practical AI tools, backed by a strong infrastructure that focuses on usability and user experience, providing immediate and tangible value.
• Collaborative AI Economy: Sahara AI creates a fair and transparent system that rewards all contributors, encouraging diverse participation and fostering innovation within the AI ecosystem.

Key Components of the Sahara AI Ecosystem:

• Sahara Blockchain: A blockchain specifically tailored for artificial intelligence, ensuring security and transparency while providing a solid foundation for transactions and managing the AI lifecycle across the ecosystem.
• AI Infrastructure: Employing the latest advanced algorithms and computing frameworks, this infrastructure supports decentralized collaborative training and services, fostering continuous advancements in AI technology.
• Sahara AI Marketplace: A decentralized platform for crowd-sourcing AI assets, bringing together a wide variety of models, agents, and datasets, while offering blockchain-based ownership verification and flexible licensing options.
• Development Tools: These user-friendly tools cater to developers and users of all skill levels, featuring the Sahara SDK and API for developers, along with no-code/low-code platforms that enable non-technical users to easily create AI solutions.
• Secure Storage: By using encrypted vaults and strong security measures, Sahara AI ensures the safety of AI assets and user data.

Together, these components form a comprehensive, user-friendly, and secure platform for collaborative AI development and deployment, facilitating the broader adoption and innovation of artificial intelligence.

How Does It Work?

The Sahara network is made up of two main layers: the execution layer and the application layer.

Execution Layer

This layer facilitates the deployment of AI, encompassing data, models, and knowledge agents, while ensuring four critical attributes: high performance, decentralization, privacy, and attribution (tracking ownership and the impact of data ownership). Individual nodes safeguard data storage, transfer data between them, and support the development of applications like Sahara Data and knowledge agents (KAs).

The execution layer employs decentralized low-rank adaptation for large language models (LoRA) to fine-tune models, ensuring high performance. It maximizes end-to-end rewards, allowing for customization and evolution while guaranteeing secure data processing and dynamic management. Additionally, it features privatized parameter handling with personalized adaptations and plug-and-play capabilities.

User privacy is upheld through privacy-preserving training and inference, allowing personal data to be securely stored on either private or decentralized public nodes, with encryption methods applied during training.

Provenance involves proof of ownership through digital watermarks and commitment schemes, as well as secure data storage and access using public key infrastructure and AES encryption. Data transmission is protected through private computation and homomorphic encryption.

Without the execution layer, the sources and attribution of individual knowledge capital could not be established. The application layer enables the democratization of the current AI economy and encourages the growth of a collaborative economy.

Application Layer

Source: Official X account of Sahara AI

The application layer is built on top of the execution layer and features several systems, including a unique built-in data marketplace for data collection, labeling, and access—something that other AI networks in Web3 currently lack.

• Sahara ID: This is Sahara’s reputation system for users, which offers provenance and security through exclusive access to their personal knowledge base. It tracks their contributions to the network in terms of achievements, income, and qualifications.
• Sahara Data Marketplace: The decentralized data marketplace integrated into Sahara has been used by various enterprise clients, such as MIT, Microsoft, USC, and Motherson Group, to help train their AI models. This marketplace includes a data toolkit (for collection, labeling, quality assurance, etc.), a distribution engine, and project management tools. We plan to expand these data operation tools for all users, whether individuals or businesses.
• Sahara Knowledge Agents (KAs): Our knowledge agents are semi-autonomous custom AI programs built on large foundational models, fine-tuned with proprietary data that remains private and secure. Users will have access to privacy and provenance toolkits, as well as a development toolkit for knowledge agents, making it easy for developers to create KAs and profit from their efforts.
• Knowledge Reconstruction: On Sahara, users can not only branch their own AI but also collaboratively develop AI within a shared ecosystem. Our model merging toolkit and knowledge agent network can integrate multiple decentralized AIs and knowledge sources. By leveraging these advanced tools, users can create richer, more interconnected AI experiences, nurturing an environment where collective intelligence can flourish.

Operating Principle

Source: Official X account of Sahara AI

Sahara is more than just a trading platform; it incentivizes the entire economic ecosystem through a process called verification staking. Using Web3 technology, it separates personal data from the training of public models, ensuring that personal information is properly attributed and that data owners receive compensation.

Participants can earn rewards in three main ways:

• By providing knowledge services, such as data, models, labeling, and data collection.
• By running nodes or offering storage to maintain the decentralized infrastructure.
• By providing verification services to ensure the network’s integrity.

The Sahara ecosystem consists of five key roles that support this economic model:

• Knowledge Consumers: They pay for model inference services, contributing funds to the system.
• Knowledge Providers: They earn compensation by offering data collection or labeling services.
• Knowledge Developers: They use these services and data to create models, earning revenue from inference services provided to consumers.
• Node Operators: They provide the necessary computing power and storage resources.
• Validators: They verify the accuracy of information and hold the system accountable for any misuse.

This structure establishes a self-sustaining ecosystem where participants can benefit from their contributions—whether operating nodes, sharing knowledge and data, verifying information, fine-tuning models, or developing autonomous agents.

Project Team and Funding Overview

The Sahara AI platform was co-founded by Professor Sean Ren from the University of Southern California and blockchain expert Tyler Zhou. Professor Ren is well-respected in AI research and innovation, with accolades including Samsung’s Annual AI Researcher Award, MIT TR35 Pioneer Innovator, and Forbes 30 Under 30. He has significantly advanced the AI field. Tyler Zhou, formerly the investment director at Binance Labs, brings extensive blockchain industry experience. Together, they have built a strong technical and industry foundation for Sahara, supported by a team of over 30 top AI contributors.

Sahara AI has raised a total of $49 million in funding. On May 20, 2024, they announced a $6 million seed round led by Polychain Capital, with participation from Sequoia Capital, Samsung Next, Nomad Capital, and others. On August 14, 2024, Sahara AI secured an additional $43 million, led by Binance Labs, Pantera Capital, and Polychain Capital, with involvement from major global investors such as Samsung, Matrix Partners, Thailand’s Kasikornbank, dao5, Alumni Ventures, Geekcartel, Nomad Capital, Mirana Ventures, and Foresight Ventures. This funding will be directed towards expanding the global team, improving AI blockchain performance, and accelerating the developer ecosystem’s growth.

Airdrop Opportunity

Sahara has not yet issued any tokens, but given its collaborative economic model, there is a strong possibility that a token will be launched in the future. For now, users can earn points by participating in activities on Galxe. You can find the activity here.

Instructions for Participation:

• Log in using your Ethereum wallet (not a social media account)
• Select “Sahara Social” and follow the on-screen instructions
• Complete the designated tasks to earn loyalty points
• If you complete all tasks, you will receive 100 loyalty points

Conclusion

Sahara AI is a decentralized blockchain platform focused on democratizing and ensuring fairness in artificial intelligence. Its unique collaborative economic model allows both individuals and businesses to engage in the development and use of AI together. While decentralized AI is still in its infancy and faces challenges like data quality, standardization, and security, it has significant potential for privacy protection, transparent ownership, and equitable distribution. As the Sahara platform evolves and potentially introduces tokens and airdrop programs, it is well-positioned for rapid growth.

Reference：

• https://x.com/SaharaLabsAI/status/1827404208058069320
• https://saharalabs.ai/blog/sahara-ai-raise-43m
• https://saharalabs.ai/blog/collaborative-ai-economy
• https://x.com/SaharaLabsAI/status/1773840384814141554
• https://foresightnews.pro/article/detail/66895
• https://www.linkedin.com/pulse/introducing-sahara-saharalabs-ai-6wqcc/