In the race to gain a competitive edge, organizations are increasingly training artificial intelligence (AI) models on sensitive data. But what if a seemingly harmless AI model became a gateway for attackers?
A malicious actor could upload a poisoned model to a public repository, and without realizing it, your team could deploy it in your environment. Once active, that model could exfiltrate your sensitive machine learning (ML) models and fine-tuned large language model (LLM) adapters. With access to these adapters, attackers could replicate your custom tuning and optimizations, exposing sensitive information embedded in fine-tuning patterns. Palo Alto Networks researchers recently uncovered two vulnerabilities in Google’s Vertex AI platform. These vulnerabilities could have allowed attackers to escalate privileges and exfiltrate models. We
Read more…
Source: Palo Alto Unit 42
Related:
- Researchers say Russian government hackers were behind attempted Poland power outage
January 23, 2026
A failed December effort to bring down parts of Poland’s energy grid was the work of Russian government hackers known for causing past energy disruptions, according to a security research firm that investigated the incident. Last week, Polish Energy Minister Milosz Motyka told reporters that the attempted cyberattack on December 29 and 30 saw hackers targeting ...
- VMware vCenter Server bug fixed in 2024 under attack today
January 23, 2026
You’ve got to keep your software updated. Some unknown miscreants are exploiting a critical VMware vCenter Server bug more than a year after Broadcom patched the flaw. The vulnerability, tracked as CVE-2024-37079, is an out-of-bounds write flaw in vCenter Server’s implementation of the DCERPC protocol that earned a 9.8 out of 10 CVSS rating. In other ...
- Data of 72 million Under Armour customers appears on the dark web
January 22, 2026
When reports first emerged in November 2025 that sportswear giant Under Armour had been hit by the Everest ransomware group, the story sounded depressingly familiar: a big brand, a huge trove of data, and a lot of unanswered questions. Since then, the narrative around what actually happened has split into two competing versions—cautious corporate statements on ...
- The Next Frontier of Runtime Assembly Attacks: Leveraging LLMs to Generate Phishing JavaScript in Real Time
January 22, 2026
Imagine visiting a webpage that looks perfectly safe. It has no malicious code, no suspicious links. Yet, within seconds, it transforms into a personalized phishing page. This isn’t merely an illusion. It’s the next frontier of web attacks where attackers use generative AI (GenAI) to build a threat that’s loaded after the victim has already visited ...
- Watering Hole Attack Targets EmEditor Users with Information-Stealing Malware
January 22, 2026
In late December 2025, EmEditor, a highly extensible and widely used text, code, and CSV editor developed by U.S.-based Emurasoft, published a security advisory warning users that its download page had been compromised. The attackers’ objective was to distribute a compromised version of the program to unsuspecting users. EmEditor has longstanding recognition within Japanese developer communities ...
- A new LinkedIn phishing scam is targeting executives online
January 21, 2026
Business executives and IT admins are being targeted by a highly sophisticated phishing attack which doesn’t happen in the email inbox but rather – on LinkedIn. Security researchers ReliaQuest said they saw a new attack that combines legitimate Python pentesting projects, DLL sideloading, and fake job ads, to infect “high-value targets” with remote access trojans ...