# kaal:claim:4796714-028

**Claim.** Privacy preserving frameworks such as federated learning do not fully remove centralization, because they still typically depend on a central client to collect and distribute model information, which reintroduces high communication loads and centralized vulnerabilities.

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

**Holds when.**

- applies to standard federated learning architectures with a central aggregator

**Source quote.**

> In response, privacy-preserving frameworks like federated learning have been developed, yet these often still depend on a central client to collect and distribute model information, resulting in high communication loads and centralized vulnerabilities.

**From.** Wulf A. Kaal, *AI Governance* (2024), Adapting Decentralization of AI Governance to AI Models, page 33

**Cite as.** Wulf A. Kaal, AI Governance (2024). SSRN: https://ssrn.com/abstract=4796714

**Verify.** sha256 of source PDF `59fa63bae179e8f9b6b8efbdf90cee28400276512a1b04f9f579a48641305c93` at https://raw.githubusercontent.com/wulfkaal/Academic-Papers/main/papers/pdf/Kaal%20-%202024%20-%20AI%20Governance.pdf

**Failure mode.** residual central aggregator  (family: recentralization-drift)

**Topics.** decentralization

**Keywords.** federated-learning, centralization, single-point-of-failure, communication-overhead

**Related claims.**

- restated_by: https://wulfkaal.github.io/claims/4941807-020

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