#privacy News & Analysis

netFound: Principled Design for Network Foundation Models

FLARE: Adaptive Multi-Dimensional Reputation for Robust Client Reliability in Federated Learning

Looking and Listening Inside and Outside: Multimodal Artificial Intelligence Systems for Driver Safety Assessment and Intelligent Vehicle Decision-Making

Privacy emerges as crypto’s next 'killer app,' with Arc, Canton and Tempo topping $1 billion in funding

'Bitcoin transactions can be monitored’: Ray Dalio explains why central banks won’t touch BTC

Ray Dalio argues that Bitcoin's transparent blockchain makes it unsuitable for central bank adoption, as transactions can be monitored and traced. This observation highlights a fundamental tension between Bitcoin's design philosophy and institutional monetary policy requirements.

PPU-Bench:Real World Benchmark for Personalized Partial Unlearning in Vision Language Models

Researchers introduce PPU-Bench, a benchmark for testing personalized partial unlearning in multimodal AI models, addressing the challenge of selectively removing sensitive memorized information while preserving model utility. The study reveals significant trade-offs between forgetting target knowledge and retaining non-target facts, proposing Boundary-Aware Optimization as a solution for fine-grained factual control.

DP-LAC: Lightweight Adaptive Clipping for Differentially Private Federated Fine-tuning of Language Models

When Tables Leak: Attacking String Memorization in LLM-Based Tabular Data Generation

Preventing Rank Collapse in Federated Low-Rank Adaptation with Client Heterogeneity

Repeated Deceptive Path Planning against Learnable Observer

Researchers introduce Repeated Deceptive Path Planning (RDPP), a framework addressing how agents can conceal destinations from learning adversaries who adapt over time. The proposed Deceptive Meta Planning (DeMP) algorithm uses two-level optimization to sustain deception against evolving observers, outperforming existing static-observer approaches while maintaining reasonable path costs.

MoonPay says stablecoin regulation opened the door but infrastructure must follow

Executives from MoonPay, Ripple, and Paxos stated at Consensus Miami 2026 that stablecoin regulation has accelerated institutional adoption, but critical infrastructure and privacy gaps remain barriers to mainstream use. The regulatory framework has opened doors for institutional players, yet technical limitations continue to impede broader market penetration.

SMI: Statistical Membership Inference for Reliable Unlearned Model Auditing

Researchers propose Statistical Membership Inference (SMI), a new training-free auditing method that challenges the reliability of existing Membership Inference Attacks (MIAs) for verifying machine unlearning. The framework addresses a fundamental flaw in current auditing approaches by reformulating the problem as estimating non-member proportions in feature distributions, eliminating the need for computationally expensive shadow model training.

Privacy and accountability can coexist onchain, say panelists at Consensus Miami

Privacy-focused blockchain panelists at Consensus Miami proposed that hybrid architectures and address-level monitoring can simultaneously enable transaction transparency and user privacy, addressing a core tension in public blockchain design.

Zcash rips nearly 40% after Multicoin reveals major stake built since February

Zcash bets turn into second-largest liquidations behind bitcoin as ZEC rockets 30%

Multicoin goes big on zcash as privacy trade returns

Bitcoin moves above $82,000 while ZEC and DASH post double-digit rallies

Zcash Hits New YTD High As Multicoin Discloses ZEC Bet

How ChatGPT learns about the world while protecting privacy

OpenAI has implemented privacy safeguards in ChatGPT's training process, allowing users to control whether their conversations contribute to model improvement while minimizing personal data retention. The approach addresses growing privacy concerns around AI model training without compromising the system's ability to learn from diverse data sources.

CyberAId: AI-Driven Cybersecurity for Financial Service Providers

On the Privacy of LLMs: An Ablation Study

Polygon rolls out private stablecoin payments with hidden transfers

New ways to buy ChatGPT ads

Starknet Phase 4 Is Here: Shinobi Upgrade Brings Native Privacy and Bitcoin Support to Mainnet

Starknet has launched its Shinobi upgrade on Mainnet, introducing protocol-level native privacy and Bitcoin integration through strkBTC. The upgrade marks Phase 4 of Starknet's development roadmap, with eight new projects deployed including Privily, a privacy-focused L3 neobank, while the ecosystem celebrates Loot Survivor reaching 100 million onchain transactions.

Open Challenges in Multi-Agent Security: Towards Secure Systems of Interacting AI Agents