Since 2023, AI technologies like ChatGPT,Google’s Gemini and OpenAI SORA started to influence our lives significantly. Despite their impact, the intricacies of their data collection and data use and the potential for inherent biases are not widely understood.
Consider an AI trained predominantly on data from a specific demographic, which could lead to unintentional biases, highlighting the importance of diverse data in development for fairness and inclusivity.
As AI becomes more intertwined with our daily lives, establishing trust in these technologies is paramount. This trust-building is akin to forming relationships with people — it grows through time spent together, through understanding and interacting. By engaging with AI, learning about its workings, and ensuring it operates on unbiased and accurate data, we foster a deeper trust in its capabilities. This approach, emphasizing transparency and diverse contributions, is essential for developing AI systems that are reliable and beneficial for everyone.
As we navigate these advancements, our goal at OORT is simple: to build an ecosystem where trust is paramount, ensuring humans and AI can thrive together.
OORT is a decentralized cloud for privacy and cost savings. By integrating global compute and storage resources, OORT empowers trustworthy AI solutions.
OORT proposes a new way of developing AI that’s open and transparent, involving key stages such as data collection, preprocessing, model design, training, fine-tuning, and deployment. By inviting everyone to participate in these steps, we aim to move beyond the opaque ‘black box’ view of AI, fostering a deeper understanding and collaboration in its creation. This approach is designed to make the AI development process clear and inclusive, building trust through active contribution and participation.
Let’s first examine the current centralized AI workflow (Figure 1). Companies like OpenAI and Google operate in a highly centralized manner, using their own data centers and performing opaque processes in their AI solutions. Users remain oblivious to how each stage of AI development is accomplished and the data used to train and fine-tune these models.
In contrast, OORT utilizes blockchain technology to redefine each stage, as illustrated in Figure 2.
Users can easily contribute data using personal devices like laptops and smartphones via the OORT platform while receiving rewards for their input.
With OORT Compute, data preprocessing occurs on user devices, such as smartphones and PCs, ensuring data privacy. Processed data is uploaded to the decentralized cloud, with its hash stored on the blockchain to ensure traceability and tamper resistance.
OORT allows and encourages community members and institutions to collaboratively design open-source AI models, enhancing innovation and accessibility.
OORT Compute offers cost-effective model training and fine-tuning, thanks to decentralized servers.
AI models are deployed on personal devices like smartphones and PCs, enabling offline use and ensuring user data privacy. This allows users to access AI tools anytime, anywhere, using their own device’s resources.
OORT’s approach to building a decentralized AI ecosystem stands out for its transparency and reliability. Unlike systems designed in isolation, OORT is built on a community-driven process — by and for everyone. This approach invites everyone not just to observe but actively contribute to AI’s progress, ensuring it’s accessible and shaped by a diverse set of voices.
OORT offers three flagship products:
Integrates global storage servers, including Seagate and Tencent Cloud, providing decentralized object storage solutions comparable to Amazon AWS S3, saving cost up to 60%.
Provides AI data collection and preprocessing services by integrating user devices, including smartphones, tablets, and PCs, ensuring data privacy while significantly reducing costs. Launching in 2024!
Provides comprehensive, trustworthy AI solutions deployable on user devices for enterprises and individuals.
In addition, OORT’s decentralized cloud infrastructure consists of two core components: OORT Edge and Olympus Protocol.
Custom-designed edge nodes dedicated to AI data computation and storage. By purchasing the edge node device — Deimos, individuals can easily become part of the OORT network as contributing nodes.
OORT’s in-house developed open-source layer-1 blockchain. Its technical details and performance theoretical proofs can be referenced in the Yellow Paper. Olympus connects devices worldwide that join the OORT network (public clouds, private clouds, smartphones, computers, etc.). Its patented proof of honesty (PoH) consensus algorithm ensures that all nodes honestly do what they are requested to do and optimizes the resource allocation of the network in a decentralized manner. Figure 3 shows the OORT architecture.
Read more about the proof of honesty consensus algorithm.
In the competitive landscape of AI + DePin (Decentralized Physical Infrastructure Networks), OORT stands out from a unique perspective. There are projects that specialize in AI data, such as Filecoin; projects focusing on AI model training, like Gensyn; and those concentrating on model algorithms, such as Bittensor.
While OORT can offer services akin to Filecoin for AI data storage and services resembling Gensyn for model training, what sets OORT apart is our emphasis on “edge AI”. That is, by connecting devices with limited computational power, such as smartphones, tablets, and PCs, OORT enables data collection, preprocessing, and model deployment transparently while safeguarding user privacy. This specialization lays a solid foundation for realizing the trustworthy AI paradigm in the future.