Overcoming Risks from Chinese GenAI Tool Usage

A recent analysis of enterprise data suggests that generative AI tools developed in China are being used extensively by employees in the US and UK, often without oversight or approval from security teams. The study, conducted by Harmonic Security, also identifies hundreds of instances in which...

On-Premise vs SaaS Data Annotation Platforms Compared

Choosing a data annotation platform? Learn when to use SaaS or on premise based on speed, cost, data privacy, and project scope...

Age verification: Child protection or privacy risk?

With governments demanding actual age verification on websites with adult content, and platforms like social media and Roblox introducing restrictions based on a user’s age, the controversy about different types of age verification and their implications is growing. Last week, Roblox announced ne...

Brave Browser Blocks Microsoft Recall from Tracking Online Activity

Brave browser now blocks Microsoft Recall by default, preventing screenshots and protecting users’ browsing history on Windows 11...

How the Solid Protocol Restores Digital Agency

The current state of digital identity is a mess. Your personal information is scattered across hundreds of locations: social media companies, IoT companies, government agencies, websites you have accounts on, and data brokers you've never heard of. These entities collect, store, and trade your...

LoRA-Leak: Membership Inference Attacks against LoRA Fine-Tuned Language Models

Language Models LMs typically adhere to a "pre-training and fine-tuning" paradigm, where a universal pre-trained model can be fine-tuned to cater to various specialized domains. Low-Rank Adaptation LoRA has gained the most widespread use in LM fine-tuning due to its lightweight computational cost...

Proton launches Lumo, a privacy-focused AI chatbot

Proton, known for its privacy focused set of services, announced the introduction of Lumo, a privacy-first Artificial Intelligence AI chatbot. It is good to know before you dive in that Proton’s chatbot has two user options that offer a very different experience. If you want Lumo to access the...

C-AAE: Compressively Anonymizing Autoencoders for Privacy-Preserving Activity Recognition in Healthcare Sensor Streams

Wearable accelerometers and gyroscopes encode fine-grained behavioural signatures that can be exploited to re-identify users, making privacy protection essential for healthcare applications. We introduce C-AAE, a compressive anonymizing autoencoder that marries an Anonymizing AutoEncoder AAE with...

Tab-MIA: a Benchmark Dataset for Membership Inference Attacks on Tabular Data in LLMs

Large language models LLMs are increasingly trained on tabular data, which, unlike unstructured text, often contains personally identifiable information PII in a highly structured and explicit format. As a result, privacy risks arise, since sensitive records can be inadvertently retained by the...

Learning-Based Privacy-Preserving Graph Publishing against Sensitive Link Inference Attacks

Publishing graph data is widely desired to enable a variety of structural analyses and downstream tasks. However, it also potentially poses severe privacy leakage, as attackers may leverage the released graph data to launch attacks and precisely infer private information such as the existence of...

CVE-2025-8039 Search terms persisted in URL bar

In some cases search terms persisted in the URL bar even after navigating away from the search page. This vulnerability was fixed in Firefox 141, Firefox ESR 140.1, Thunderbird 141, and Thunderbird 140.1...

Towards Trustworthy AI: Secure Deepfake Detection Using CNNs and Zero-Knowledge Proofs

In the era of synthetic media, deepfake manipulations pose a significant threat to information integrity. To address this challenge, we propose TrustDefender, a two-stage framework comprising i a lightweight convolutional neural network CNN that detects deepfake imagery in real-time extended...

The vulnerability of the Aggregate Term Handler component in the SQLite database management system allows a attacker to compromise privacy, integrity, and accessibility.

The vulnerability of the Aggregate Term Handler component in the SQLite database management system is related to numerical truncation errors. Exploiting this vulnerability could allow an attacker to compromise privacy, integrity, and accessibility of the data...

Mozilla -- Persisted search terms in the URL bar

[email protected] reports: In some cases search terms persisted in the URL bar even after navigating away from the search page...

AUTOPSY: a Framework for Tackling Privacy Challenges in the Automotive Industry

With the General Data Protection Regulation GDPR in place, all domains have to ensure compliance with privacy legislation. However, compliance does not necessarily result in a privacy-friendly system as for example getting users' consent to process their data does not improve the...

A week in security (July 14 – July 20)

Last week on Malwarebytes Labs: Meta execs pay the pain away with $8 billion privacy settlement Adoption agency leaks over a million records Meta AI chatbot bug could have allowed anyone to see private conversations WeTransfer walks back clause that said it would train AI on your files Chrome fix...

Multi-Stage Prompt Inference Attacks on Enterprise LLM Systems

Large Language Models LLMs deployed in enterprise settings e.g., as Microsoft 365 Copilot face novel security challenges. One critical threat is prompt inference attacks: adversaries chain together seemingly benign prompts to gradually extract confidential data. In this paper, we present a...

Optimizing Canaries for Privacy Auditing with Metagradient Descent

In this work we study black-box privacy auditing, where the goal is to lower bound the privacy parameter of a differentially private learning algorithm using only the algorithm's outputs i.e., final trained model. For DP-SGD the most successful method for training differentially private deep...

DP2Guard: a Lightweight and Byzantine-Robust Privacy-Preserving Federated Learning Scheme for Industrial IoT

Privacy-Preserving Federated Learning PPFL has emerged as a secure distributed Machine Learning ML paradigm that aggregates locally trained gradients without exposing raw data. To defend against model poisoning threats, several robustness-enhanced PPFL schemes have been proposed by integrating...

Metaverse Security and Privacy Research: a Systematic Review

The rapid growth of metaverse technologies, including virtual worlds, augmented reality, and lifelogging, has accelerated their adoption across diverse domains. This rise exposes users to significant new security and privacy challenges due to sociotechnical complexity, pervasive connectivity, and...