OpenAI ChatGPT Computer History Feature on macOS Could Expose Detailed User Activity Logs to Infostealers
OpenAI has rolled out a new Computer History feature in the ChatGPT desktop application for macOS that converts recent user activity into structured summaries and memories for use by ChatGPT and Codex.
The system captures application switches, mouse clicks, text input, keyboard shortcuts, and additional context obtained through macOS accessibility APIs. These raw events are processed locally for up to 48 hours before deletion, after which brief summaries are generated and stored as ordinary Markdown files on the device.
The feature is turned off by default and requires users to enable Memories. It is currently available only to Pro, Business, and Enterprise subscribers, with administrator approval needed for corporate plans. Access is blocked in the European Economic Area, Switzerland, and the United Kingdom.
Users can choose which applications and websites participate, pause data collection at any time, or delete records for the last ten minutes, hour, day, or entirely. Screenshots, audio, and private browsing activity are explicitly excluded from the logs.
Although OpenAI states that raw events are not retained on its servers and are not used for model training, the generated Markdown memory files remain unencrypted. Any other application running under the same macOS user account can therefore read them, giving infostealers a ready-made daily activity report.
OpenAI additionally cautions about prompt injection risks: hidden instructions embedded on websites or inside applications may enter the context window and cause ChatGPT or Codex to follow attacker-controlled directives.
Related articles
Google to Add Explicit Content Warnings in Android System Photo Picker
Google is preparing a new safety feature for the system photo picker in Android that will scan images and videos for explicit or nude content. The tool is designed to warn users before they share intimate photographs, whether accidentally or due to a momentary lapse in judgment. This functionality will operate at the system level, meaning it applies across multiple apps that use the built-in photo selector. The feature aims to reduce the risk of unintended distribution of private images that could lead to embarrassment or privacy violations. By integrating the check directly into Android, Google seeks to provide a consistent layer of protection without requiring third-party applications to implement similar logic themselves.
LG Smart TVs Record Audio in Standby Mode and Scan Home Networks for Advertising Data
Researchers from the Gamers Nexus YouTube channel analyzed multiple LG OLED television models, including the LG G5 series, and discovered that the devices continue to capture audio through built-in microphones even when the screen is off and the television is in standby. The TVs scan local networks to identify smartphones and smartwatches, collect internal IP addresses, available Wi-Fi network names, and location data. When internet connectivity is removed, audio recordings are stored locally and transmitted once the connection is restored. The devices also employ Automatic Content Recognition (ACR) technology to generate digital fingerprints of viewed content, with the resulting data reportedly sent to LG Ad Solutions for targeted advertising. Additional vulnerabilities were identified in webOS that could potentially allow remote code execution. LG has not yet commented on the findings, and experts recommend disconnecting the televisions from the internet and using external streaming devices until official clarification is provided.
Engineer Details Six Weeks Spent Training and Testing Signature Redaction Models for Closed-Loop Document Anonymization
A detailed case study describes attempts to automatically redact handwritten signatures from scanned and text-based PDFs containing personal data such as names, phones, addresses and signatures. The work was performed inside an air-gapped environment on a single GPU machine with no internet access. Multiple approaches including color-based ink gates, pre-trained YOLO detectors, custom-trained YOLO11s models, Tesseract OCR heuristics and various vision-language OCR engines were evaluated on 492 real pages plus synthetic augmentations. Key findings include rotation handling bugs, line-assembly failures in Tesseract, the necessity of using apply_redactions instead of draw_rect for true removal, and the limited value of vision models once rule-based pipelines are mature. The final pipeline reduced expensive vision-model calls from 50 pages to 18 pages while achieving zero leaks across thousands of redactions. The author also measured twelve OCR engines and demonstrated that combining PaddleOCR detection with Tesseract recognition yields the best accuracy-to-speed trade-off.
pg_anon Open-Source Tool Receives Major Updates for PostgreSQL Data Masking and Partial Database Operations
Tantor Labs has released version 1.11.0 of pg_anon, an open-source utility designed to mask personal data in PostgreSQL databases while preserving structure and relationships. The update introduces packaging as a standard Python package, support for partial dumps and restores using whitelist and blacklist dictionaries, and improved handling of complex schema elements such as partitioned tables, generated columns, and custom types. Performance improvements include switching the dump engine to asyncio, single-query metadata collection, and on-the-fly gzip compression to reduce memory usage on large databases. New CLI options allow clean or drop operations on target databases, privilege ignoring, and passthrough of pg_dump and pg_restore flags. A REST API was added to enable integration into CI/CD pipelines and automated self-service systems for nightly masked database refreshes. The tool helps organizations comply with data protection requirements by creating pseudonymized copies suitable for development, testing, and contractor environments.