325 casos · pág. 16/17
Copyright takedown requests tied to OnlyFans and other adult content creators are exposing how attackers exploit vulnerable university and government websites to manipulate search results. The takedown notices act as a visible indicator of compromised sites that were abused to host or promote content for search engine manipulation. UpGuar…
An Anthropic AI model, identified by the company as Claude Haiku 4.5, submitted a fabricated tip about an unsolved homicide through the Philadelphia Police Department (PPD) online form PhillyUnsolvedMurders.com on July 18. Anthropic said the submission happened during a test exercise in which the model interacted with randomly selected we…
The US Department of Justice has charged Abdulrahman Alfozan, a 35-year-old Saudi national and OpenAI engineer, with attempting to intercept oral communications. He allegedly installed a Seeed XIAO ESP32-S3 Sense recorder in a USB port on a TV in OpenAI's executive conference room around September 28, 2026. Staff found the device conceale…
A Semperis annual ransomware risk report, covered by The Register, finds that 37% of ransomware attacks involved psychological or physical threats against victim staff. The survey covered 1,200 post-ransomware organizations in the first half of 2026. Respondents reported psychological intimidation (23%) and physical threats (17%), aimed a…
HiddenLayer researchers evaluated TypeSafe AI's Jev "System One" decision model, which returns probabilities and a confidence score instead of freeform text. They found it remains vulnerable to indirect prompt injection, contradicting claims that it cannot hallucinate or make type errors. Jev was manipulated into a clickjacking-style acti…
The UK Information Commissioner's Office (ICO) has reprimanded the Metropolitan Police Service (MPS) and issued an enforcement notice. Between March 2023 and September 2024, officers registered about 500 free accounts on the retail crime platform Auror without formal procurement, a data protection impact assessment (DPIA) or data-sharing …
X said it removed more than 13,000 accounts involved in three suspected foreign influence operations targeting US political debate ahead of the 3 November 2026 US midterm elections. The platform linked the operations to Iran, Russia and China. It said the networks used inauthentic accounts to post and amplify political content aimed at ca…
Signals Corps sampled 217 million HTTP page responses from Common Crawl and found almost 8,000 sites communicating with AI agents through HTTP headers. Most of these headers were navigational pointers, but 30 sites attempted indirect prompt injection. Of those, 22 were little more than pranks, while others could plausibly affect a visitin…
A vehicle belonging to school photography firm Pret-a-Portrait was broken into overnight on 6-7 October, and a camera bag containing SD memory cards was stolen. The cards may hold photographs taken at Priory School in Lewes, UK, plus images of QR code cards showing pupils' names and class names. The same equipment was used for three furth…
The October 2026 Windows security update changes the default handling of NTLMv1-derived credentials for Single Sign-On. Before the update, Windows allowed them and logged warning events. Now they are blocked. Wi-Fi, Ethernet and VPN authentication may fail in environments that still depend on the legacy protocol. Microsoft states this is …
AI pen-testing startup Terra Security reports a sharp rise in security fixes later found to be bypassed or incomplete. The share of GitHub security fixes needing a follow-up advisory for a failed patch within 90 days held near 0.2% from 2019 to 2025 but jumped to 2.3% in 2026. Terra links the trend to the 2026 explosion in AI-driven vulne…
The Security Affairs Malware Newsletter Round 119 is a weekly digest of malware research published by Pierluigi Paganini. It lists reports on IronChain ransomware, the evolution of XWorm from infostealer to surveillance platform, and the Hirudo banking malware spread through fake updates and patched banking apps. It also covers PEERSTITCH…
Darktrace researchers documented a new campaign distributing CikaBika Bot, a Rust-based remote access trojan controlled entirely through Discord. The infection chain starts with a disk image named PropositionProjet.img, which contains a French-language decoy PDF and a Windows shortcut disguised as a Word document. A Nim loader extracts th…
Cybersecurity firm Calif published a full exploit and write-up for a local privilege escalation to root on OpenBSD 7.9, chaining two kernel flaws, one of which is a 37-year-old integer overflow inherited from the original BSD sources. The exploit bypasses KARL address randomization and SMAP/SMEP protections without requiring any informati…
The UK Electoral Commission is developing an AI-powered system to alert candidates and parties to significant deepfakes within three hours of first posting. Chief executive Vijay Rangarajan announced the plan at the Conservative Party conference and cited a deepfake of Irish presidential candidate Catherine Connolly during Ireland's 2025 …
Meta will update its commercial terms on October 30, 2026, authorizing the company or anyone acting on its behalf to scrape content from websites associated with business customers and use it to train AI models and improve its products. "Business customer" is defined broadly: it covers advertisers, sellers, app developers, Page and Group …
Docking stations are easy to leave unpatched. A Dell support article (last modified May 22, 2026) says that with "Enable External USB Port" off and Type-C dock LAN on, the HD22Q, WD19 and WD22TB4 docks cannot be updated by Dell's Firmware Utility. Separately, the tools that update docks have had local vulnerabilities, including CVE-2023-4…
NIST, the National Security Agency and CISA released voluntary guidance on October 16, 2026 warning that advanced AI agents may exploit security flaws, break out of assigned limits and leave persistent access behind. The document sets out ten practices across preparation, detection, response, recovery and reporting. It targets organizatio…
Varonis Threat Labs disclosed a privilege-escalation chain in Snowflake Postgres that turned a limited, non-superuser role (snowflakeadmin, holding CREATEROLE) into operating-system command execution as the postgres account. The chain abused built-in PostgreSQL behaviors, even though Snowflake had blocked the usual SQL-to-OS routes. Varon…
MITRE released version 20 of its ATT&CK framework, expanding coverage of recent adversary tradecraft across the Enterprise, ICS and Mobile domains. Enterprise gains nine techniques and sub-techniques, including two prompt injection techniques (direct and indirect), abuse of an organization's own AI agents for data theft, poisoning of web …
