A recent study published on arXiv highlights advancements in federated learning, a technique known for enhancing privacy during the training of machine learning models by keeping sensitive data decentralized (F1). The study addresses the inherent risks of exposing private information through model updates and proposes the use of Quadratic Unconstrained Binary Optimization (QUBO) as a method to mitigate these privacy risks (F2). While the claim that QUBO improves federated learning privacy is currently unverified (A1), the research suggests that integrating QUBO could strengthen the protection of sensitive data in federated learning frameworks. This development aligns with ongoing efforts to balance data utility and privacy in distributed machine learning environments. The study contributes to the broader discourse on privacy-preserving AI, emphasizing the importance of innovative optimization techniques in safeguarding user information. Further validation and empirical testing will be necessary to confirm the effectiveness of QUBO in this context.