Adaptive Model Inversion Attacks Generalize a Privacy-Robustness Tradeoff

In this paper, we show that standard evaluations of high-resolution Model Inversion Attacks MIAs significantly underestimate training-data privacy leakage. State-of-the-art privacy defenses, standard training techniques such as MixUp and Adversarial Training, and undefended models all leak traini...

AgentDoxx: Agentic Re-Identification of Anonymized Text with Web Search

As Large Language Models LLMs gain tool use capabilities such as web search, they can retrieve and cross-reference public information, creating privacy risks beyond memorization. One manifestation is re-identification: linking an anonymized interview transcript to a named individual. Yet without...

The Poisoned Conversation: Privacy-Leaking Watermarks in Unified Multimodal Models

Multimodal models are increasingly shifting toward unified architectures that understand and generate text, images, and other modalities within a shared conversational context. This design enables fluid interaction across modalities, but it also changes the privacy threat model: Information...

Muse Creates Detailed Profiles of All Your Friends and Family

Millions have downloaded Meta’s AI agent Muse. But getting it to do your bidding comes with privacy costs...

EUVD-2026-91881

The Ultimate Member – User Profile, Registration, Login, Member Directory, Content Restriction & Membership Plugin plugin for WordPress is vulnerable to authorization bypass in all versions up to, and including, 2.13.1 This is due to the plugin not properly verifying that a user is authorized to...

CVE-2026-93428

The Ultimate Member plugin for WordPress is vulnerable to an authorization bypass in all versions up to and including 2.13.1 . The root cause is the plugin's failure to properly verify user authorization when accessing the publicly available wp_ajax_nopriv_um_get_members endpoint. Specifically, t...

CorrectGuard: Eyes-Off Correctness Estimation for Black-Box Security Guardrails

AI services increasingly rely on black-box security guardrails, yet privacy-preserving model auditing regimes often cannot measure how well these systems perform in both a human eyes-off production setting, which disallows human inspection of user input, and a machine eyes-off setting, which...

CVE-2026-104051

PictShare versions prior to 3.7.1 are vulnerable to an information disclosure vulnerability via the API::info() endpoint. The root cause is the lack of a field whitelist, allowing unauthenticated attackers to retrieve complete raw metadata objects. By using a publicly visible file hash, an attack...

OverAct: Measuring and Mitigating Proactive Over-Authorization in LLM Tool-Calling Agents

LLM agents with tool-calling capabilities can access external services and private user data, but they may retrieve more information than a user's request explicitly requires. We study this behavior in structured tool-calling agents and term it proactive over-authorization. This setting differs...

Sleeping Secrets: How Fine-Tuning Reawakens Privacy Risks in Language Models

Beyond adapting Large Language Models LLMs to specialized applications, fine-tuning has recently been shown to recover private information that is no longer accessible through direct queries. Previous fine-tuning recovery attacks, however, require genuine private supervision drawn from the same...

Mitigating Private Data Leakage in LLMs with Whiteout

Modern large language models LLMs are trained on massive, largely unfiltered datasets, including content scraped from nearly every accessible website and user inputs. As a result, LLMs often memorize and reproduce personally sensitive information PSI such as birth dates, phone numbers, and home...

"I'm Trying Not to Get Hacked:" How Adults with Intellectual and Developmental Disabilities Navigate Security and Privacy Notifications

Security and privacy notifications, such as login alerts, spam email warnings, and cookie consent requests, play a critical role in shaping users' responses to digital risks. Yet most notifications overlook cognitive accessibility, limiting their effectiveness for people with intellectual and...

Your car’s app could be telling Big Tech who you are and where you go

A study testing 21 cars from 19 brands and 30 companion apps found connections to advertising and tracking companies, and evidence that some apps shared sensitive personal data. Modern cars can unlock remotely, route around traffic, stream entertainment, summon roadside help, and cool or heat the...

Link Inference Attack on Privacy-Preserving Knowledge Graphs

Knowledge Graphs KGs are widely used to store and share structured information across sensitive domains such as healthcare, fi- nance, and social networks. A common privacy practice is to delete sen- sitive relations before publishing the graph, under the assumption that removing edges is...

PrivCert: Certifying Statement Support under Differential Privacy

Differentially private DP text generation can protect individual records, but privacy alone does not specify what evidence a released statement carries about the underlying data. We identify this as an evidence gap: a private report may contain plausible claims without indicating whether they are...

CVE-2026-97472

A flaw was found in the Linux kernel's IPv6 networking subsystem. An attacker on the local network can cause a denial of service to IPv6 temporary address generation by sending router advertisements that temporarily deprecate and then restore a network prefix. Because the kernel improperly...

Meta’s Muse sent a Facebook Marketplace buyer to a seller’s home

A Facebook Marketplace buyer arrived at a seller’s apartment to collect a keyboard. The seller wasn’t home and didn’t know anyone was coming. Meta’s AI assistant Muse had handled the conversation, shared his address, and arranged the visit without telling him. We often write about AI misalignment...

Humans are reviewing Copilot users’ bizarre and abusive image-editing requests

A reminder for anyone who thinks their AI chats are private: A human may see what they upload. A report from 404 Media this week describes contractors reviewing Copilot users’ photos, editing requests, and the images Copilot produces. Some of what they see is disturbing. The third-party companies...

Privacy in Personalized AI Is a System Property, Not Just a Model Property

In personalized AI applications, such as conversational assistants and recommender systems, users interact not with models in isolation but with broader systems that access, infer, and reuse user information across components and over time. While such use of user information is integral to...

Making the Most of Leftovers: Improved Privacy Amplification for Quantum Key Distribution

The amount of secret key that can be obtained from a quantum key distribution run depends on both the physically observed error rates and the mathematical bounds used to certify security. For finite datasets, conservative bounds force users to discard a substantial fraction of the potentially...