Codatta and Alibaba Cloud Collaborate to Advance Community AI Innovation
SINGAPORE — July 18, 2025 — Codatta, a Web3-native platform for collaborative data contribution, is joining forces with Alibaba Cloud, the leading cloud infrastructure and AI service provider, to explore new frontiers in community-driven artificial intelligence. The collaboration brings together Codatta’s decentralized contributor network and Qwen’s large-scale AI models to deliver tailored AI solutions across industries, with a focus on quality, transparency, and real-world application, which is truly inspiring.
This collaboration is centered around integrating the Codatta protocol with Alibaba Cloud to deliver a more complete and human-augmented AI development experience.
At the technical level, the collaboration connects with Alibaba’s open-source model family — a growing suite of language and multimodal models — and its PAI platform (Platform for AI), which is used for model fine-tuning, deployment, and end-to-end AI development. By combining these components, the project creates a full-stack AI solution: from data collection and human annotation, to model training and inference, all within the Alibaba Cloud ecosystem.
This makes it easier for developers and enterprises to build more capable, fine-tuned AI systems with human-in-the-loop processes, without needing to leave the Alibaba Cloud environment.
Building Better Models with Community-Contributed Knowledge
Codatta will fine-tune Qwen’s large language and vision models (LLMs and LVMs) using domain-specific datasets contributed by its global network. These datasets are carefully curated and evaluated through human-in-the-loop systems. Codatta’s expert contributor base also provides human intelligence services that play a direct role in shaping model behavior. It offers deeper insight, relevance, and safety across applications. The result is domain-specific AI tailored to real-world needs in finance, healthcare, e-commerce, and more.
Codatta’s contributor network brings together well-curated and diverse datasets and expert human input. This combination enhances the performance of Qwen’s base models, enabling developers to create AI tools that are more precise and better suited to real-world applications. The outcome is domain-specific AI shaped by people who know, and, which is very important, want to know the field.
Codatta is also using Qwen-based AI agents to help automate data labeling and model evaluation. These agents can annotate large batches of content quickly, generating first-draft labels that reduce time and cost. This agent-first process is tightly integrated with Codatta’s human expert review pipeline. It forms a hybrid AI + human workflow. Human contributors then step in to review, correct, and refine the entire output.
This shared workflow allows startups and enterprise teams to build and fine-tune models more efficiently without compromising on quality. Automation handles the bulk work, while human reviewers ensure precision and handle complex or ambiguous cases.
The partnership also covers synthetic data generation for situations where real-world information is limited, difficult to obtain, or too sensitive to use directly. Using Qwen’s generative capabilities, the teams can create artificial datasets that reflect the structure and patterns of real data.
This approach is especially valuable in areas like robotics, medical research, and cybersecurity, where access to reliable data is often restricted by safety rules, compliance requirements, or high costs. It allows developers to train stronger models, protect sensitive information and meet ethical standards.
Contributors Get a Real Stake in AI Development
Codatta’s system gives contributors actual ownership of the training datasets they help create. That means when those datasets power real-world applications, the contributors earn ongoing royalties tied to their involvement.
This changes the typical AI workflow, where people provide information but see no benefit from how it is actually used. With Codatta, contributors are part of the value chain. The platform includes built-in royalty sharing, scoring systems, and token rewards to recognize valuable work and ensure contributors are fairly compensated.
A Shared Commitment to Open and Transparent AI
Alibaba Cloud has long supported open-source AI by releasing over 100 models for public use. This collaboration with Codatta builds on that vision by combining open-source technology with a community-driven model for data contribution.
“Fine-tuning open foundation models with community-powered data is a game-changer for building trustworthy AI,” said Yi Zhang, Co-founder and CEO of Codatta. “This collaboration shows that decentralized contributor networks like ours can drive innovation and make sure the people who share their knowledge and time are rewarded fairly.”
Both Codatta and Alibaba Cloud see this partnership as a move toward more inclusive and community-led AI development. Their shared goal: to lower the barriers for builders and give contributors a real stake in the AI they help shape.
For more information or to get involved, visit codatta.io or follow @codatta_io on X.
Stay Updated:
Website || X (Twitter) || Telegram || Discord || Links
