# kaal:claim:4855607-006

**Claim.** Federated learning does not eliminate privacy risk: because gradients and partial parameters are transmitted, the system remains vulnerable to attacks that leak data, and this vulnerability together with communication overhead is a significant hurdle to deployment.

**Type.** failure  **Support.** evidenced

**Holds when.**

- federated learning schemes that share gradients or partial parameters with a server

**Source quote.**

> Privacy concerns, reliance on batch-by-batch updates, vulnerability to data leaks caused by attacks due to the transfer of gradients and partial parameters, and communication overhead are significant hurdles that need to be overcome for the successful deployment of FL.

**From.** Wulf A. Kaal, *How AI Models are Optimized Through Web3 Governance* (2024), Model Overview: Federated Machine Learning Models, page 16

**Cite as.** Wulf A. Kaal, How AI Models are Optimized Through Web3 Governance (2024). SSRN: https://ssrn.com/abstract=4855607

**Verify.** sha256 of source PDF `eb0b3e62374b45a8fa888c6bde9725e606bcb46cf4b5e74a6e851d9f25099113` at https://raw.githubusercontent.com/wulfkaal/Academic-Papers/main/papers/pdf/Kaal%20-%202024%20-%20How%20AI%20Models%20are%20Optimized%20Through%20Web3%20Governance.pdf

**Failure mode.** Gradient leakage in federated learning  (family: privacy-and-surveillance-risk)

**Topics.** consensus-and-security

**Keywords.** federated-learning, privacy-leakage, gradient-attacks, data-leaks, deployment-barriers

**Related claims.**

- extends: https://wulfkaal.github.io/claims/4796714-009
- restated_by: https://wulfkaal.github.io/claims/4941807-012
- extended_by: https://wulfkaal.github.io/claims/4941807-020
- contested_by: https://wulfkaal.github.io/claims/5095633-017

**Canonical form.** This markdown file is the canonical hashed representation of the claim. Its sha256 is the content hash used for attestation.