A recent highlight is FLock.io choosing the Walrus protocol as its core decentralized data layer for building privacy-preserving AI model training platforms.

FLock.io is a decentralized learning-based AI platform that enables the community to collaboratively train AI models through federated learning, without centralizing data on any centralized server. The key to this model is that both the data itself and the training process need to be secure and reliable.

@Walrus 🦭/acc as the data foundation layer not only provides decentralized storage but also incorporates Seal's access control mechanism, allowing sensitive data to be securely encrypted and stored, with access permissions controlled through smart contracts. This way, AI model training data can be read and used by multiple nodes while maintaining privacy.

This integration greatly reduces the risks of data leakage and centralized control faced by traditional AI platforms. For developers who wish to avoid hosting data with large tech companies, this is a revolutionary change.

At the same time, it also provides a more realistic foundation for the implementation of decentralized AI: AI training datasets tend to be very large, while Walrus can store this data at a lower cost and with higher efficiency. In the future, we can expect to see more AI projects deploying data layers directly on Walrus, rather than relying on traditional databases or cloud storage services.

#walrus $WAL