#privacy News & Analysis

AI Evasion and Impersonation Attacks on Facial Re-Identification with Activation Map Explanations

Researchers developed a novel framework for generating adversarial patches that can fool facial recognition systems through both evasion and impersonation attacks. The method reduces facial recognition accuracy from 90% to 0.4% in white-box settings and demonstrates strong cross-model generalization, highlighting critical vulnerabilities in surveillance systems.

Membership Inference for Contrastive Pre-training Models with Text-only PII Queries

Researchers developed UMID, a new text-only auditing framework to detect if personally identifiable information was memorized during training of multimodal AI models like CLIP and CLAP. The method significantly improves efficiency and effectiveness of membership inference attacks while maintaining privacy constraints.

VisualLeakBench: Auditing the Fragility of Large Vision-Language Models against PII Leakage and Social Engineering

Researchers introduced VisualLeakBench, a new evaluation suite that tests Large Vision-Language Models (LVLMs) for vulnerabilities to privacy attacks through visual inputs. The study found significant weaknesses in frontier AI systems like GPT-5.2, Claude-4, Gemini-3 Flash, and Grok-4, with Claude-4 showing the highest PII leakage rate at 74.4% despite having strong OCR attack resistance.

Learnability and Privacy Vulnerability are Entangled in a Few Critical Weights

Researchers discovered that privacy vulnerabilities in neural networks exist in only a small fraction of weights, but these same weights are critical for model performance. They developed a new approach that preserves privacy by rewinding and fine-tuning only these critical weights instead of retraining entire networks, maintaining utility while defending against membership inference attacks.

Ethereum Users Warned as USDT Dust Attacks Jump 612%

Ethereum dust attacks involving USDT have surged by 612%, with USDC transfers also experiencing significant increases. Researchers are issuing warnings to Ethereum users about this spike in malicious small-value transactions designed to compromise privacy and track wallet addresses.

XRP on the Verge of Privacy Breakthrough: Top XRPL Contributor Details How This Feature Will Be Realized

XRP Ledger is developing privacy features through zero-knowledge proofs and a new token standard, according to top XRPL contributor Vet. The implementation aims to provide bank-grade confidentiality while maintaining regulatory compliance through auditability features.

Ethereum Foundation publishes formal mandate to hard‑lock censorship resistance and privacy

Explainable LLM Unlearning Through Reasoning

Researchers introduce Targeted Reasoning Unlearning (TRU), a new method for removing specific knowledge from large language models while preserving general capabilities. The approach uses reasoning-based targets to guide the unlearning process, addressing issues with previous gradient ascent methods that caused unintended capability degradation.

Jay Clayton’s Crypto War

Jay Clayton, former Trump-appointed SEC chair and predecessor to Gary Gensler, continues aggressive prosecution efforts targeting cryptocurrency privacy developers. This ongoing regulatory enforcement represents a sustained campaign against privacy-focused crypto projects and their developers.

Prosecutors push to retry Tornado Cash founder even after Washington said crypto mixers have legal uses

The U.S. Treasury acknowledged that crypto mixers have legitimate uses for privacy protection, while SDNY prosecutors simultaneously moved to retry Tornado Cash co-founder Roman Storm. This creates conflicting signals from Washington about the legal status of cryptocurrency mixing services.

Meissa: Multi-modal Medical Agentic Intelligence

Researchers have developed Meissa, a lightweight 4B-parameter medical AI model that brings advanced agentic capabilities offline for healthcare applications. The system matches frontier models like GPT in medical benchmarks while operating with 25x fewer parameters and 22x lower latency, addressing privacy and cost concerns in clinical settings.

Quantum Computing Isn't Just Coming for Bitcoin—It Threatens Messaging Apps Too

Quantum computing advances pose a significant threat to encrypted messaging applications through 'harvest now, decrypt later' attacks, where adversaries collect encrypted data today to decrypt it once quantum computers become capable enough. This risk extends beyond Bitcoin and cryptocurrencies to affect everyday communication security.

Starknet to deploy STRK20 framework enabling privacy-focused stablecoins and other assets

Starknet is developing the STRK20 framework to enable privacy-focused stablecoins and other digital assets with built-in confidentiality features. This framework represents a significant step toward enhanced privacy in decentralized finance applications on the Starknet network.

Trump's DOJ Seeks October Retrial for Tornado Cash Developer Roman Storm

The Department of Justice is pushing for an October retrial of Tornado Cash developer Roman Storm, despite the U.S. Treasury's recent acknowledgment that cryptocurrency mixers can have legitimate use cases. This represents a continued legal battle over privacy-focused blockchain tools and their regulatory treatment.

'Privacy Is Coming for XRP': Top Contributor Confirms

XRP is set to receive privacy features through amendment XLS-372, which introduces Confidential Multi-Purpose Tokens (MPTs) to the XRP Ledger. This development follows a recent US Treasury policy shift regarding blockchain privacy, as confirmed by top XRPL contributor Vet.

The Download: murky AI surveillance laws, and the White House cracks down on defiant labs

A public dispute between the Department of Defense and AI company Anthropic has highlighted unresolved questions about the Pentagon's authority to use AI for surveillance of American citizens. The conflict raises important legal and constitutional issues regarding AI surveillance capabilities and oversight.

Treasury tells Congress mixers have valid privacy uses, recommends ‘hold law’ for suspicious crypto

The U.S. Treasury told Congress that cryptocurrency mixers have legitimate privacy uses while recommending a 'hold law' for suspicious crypto transactions. Treasury data reveals that over $1.6 billion from mixing services has flowed into crypto bridges since May 2020, highlighting the scale of mixer usage in DeFi.

US Treasury report acknowledges legitimate uses of crypto mixers

The US Treasury has released a report to Congress acknowledging legitimate uses of cryptocurrency mixers, marking a notable shift in regulatory perspective. The report was commissioned under the GENIUS stablecoin regulatory framework directives.

Is the Pentagon allowed to surveil Americans with AI?

A public dispute between the Pentagon and AI company Anthropic has highlighted unresolved legal questions about whether the US government can conduct mass surveillance on Americans using AI technology. The controversy emerges more than a decade after Edward Snowden's revelations about NSA bulk data collection, indicating ongoing ambiguity in surveillance laws.

Workers report watching Ray-Ban Meta-shot footage of people using the bathroom

Meta faces accusations of concealing privacy facts about Ray-Ban smart glasses after workers reported viewing footage of people in bathrooms. The allegations raise serious concerns about user privacy and data handling practices for wearable AI devices.

Meta sued over AI smart glasses’ privacy concerns, after workers reviewed nudity, sex, and other footage

Meta faces a lawsuit over privacy concerns regarding its AI smart glasses, with allegations that the company's marketing promised user control while subcontractors were actually reviewing customer footage including sensitive content. The legal action centers on discrepancies between Meta's privacy promises and actual data handling practices.

Meta’s AI glasses reportedly send sensitive footage to human reviewers in Kenya

Meta's AI-powered smart glasses are reportedly sending sensitive footage including intimate moments to human reviewers in Kenya, according to a Swedish investigation. A class action lawsuit has emerged accusing Meta of violating privacy laws and false advertising regarding their privacy claims.

AI tools can unmask anonymous accounts

Researchers from ETH Zurich, Anthropic, and other institutions have developed AI tools that can unmask anonymous online accounts by analyzing behavioral patterns and information across platforms. The study, which has not yet been peer reviewed, suggests AI agents can identify users behind pseudonymous accounts on platforms like Reddit, X, and Glassdoor.

Chamath Palihapitiya questions bitcoin’s role as central bank reserve asset

Billionaire venture capitalist Chamath Palihapitiya has raised concerns about Bitcoin's viability as a central bank reserve asset, citing privacy and fungibility limitations. His comments come amid ongoing debate about corporate Bitcoin adoption strategies, particularly referencing MicroStrategy's substantial Bitcoin holdings.

PRIVATEEDIT: A Privacy-Preserving Pipeline for Face-Centric Generative Image Editing

Researchers have developed PRIVATEEDIT, a privacy-preserving pipeline for face-centric image editing that keeps biometric data on-device rather than uploading to third-party services. The system uses local segmentation and masking to separate identity-sensitive regions from editable content, allowing high-quality editing while maintaining user control over facial data.