Large language models (LLMs) and AI agents are becoming deeply integrated into web browsers, search engines and automated content-processing pipelines. While these integrations can expand functionality, they also introduce a new and largely underexplored attack surface.
One particularly concerning class of threats is indirect prompt injection (IDPI), in which adversaries embed hidden or manipulated instructions within website content that is later ingested by an LLM. This article shares in-the-wild observations from our telemetry, including our first observed case of AI-based ad review evasion.
Read more…
Source: Palo Alto Unit 42
Sign up for the Cyber Security Review Newsletter
The latest cyber security news and insights delivered right to your inbox
Related:
- MyCert: Malaysia data breaches up 29% in Q1 2025
June 11, 2025
The Malaysia Computer Emergency Response Team (MyCert) reported an increase in data breach incidents in Malaysia in the first quarter of the year. “Data breach incidents are growing in Malaysia with a nearly 29% increase this quarter, underscoring the need for better security measures to ensure national security and public trust,” said MyCert. According to its ...
- US government’s vaccine website defaced with AI-generated content
June 11, 2025
A U.S. government website designed to inform the public about vaccines has been defaced and now hosts apparently AI-generated spam. The domain, which belongs to the U.S. Department of Health and Human Services (HHS), appears to have been hosting the same kind of content — mostly gay-themed and LGBTQ+ posts — since at least May 12, ...
- Toxic trend: Another malware threat targets DeepSeek
June 11, 2025
DeepSeek-R1 is one of the most popular LLMs right now. Users of all experience levels look for chatbot websites on search engines, and threat actors have started abusing the popularity of LLMs. kaspersky previously reported attacks with malware being spread under the guise of DeepSeek to attract victims. The malicious domains spread through X posts and ...
- BlackSuit Continues Social Engineering Attacks in Wake of Black Basta’s Internal Conflict
June 10, 2025
There has been a significant decrease in social engineering attacks linked to the Black Basta ransomware group since late December 2024. This lapse also included the leaked Black Basta chat logs in February 2025, indicating internal conflict within the group. Despite this, Rapid7 has observed sustained social engineering attacks. Evidence now suggests that BlackSuit affiliates have ...
- Patch Tuesday – June 2025
June 10, 2025
Microsoft is addressing 67 vulnerabilities this June 2025 Patch Tuesday. Microsoft has evidence of in-the-wild exploitation for just one of the vulnerabilities published today, and that is reflected in CISA KEV. Separately, Microsoft is aware of existing public disclosure for one other freshly published vulnerability. Microsoft’s luck holds for a ninth consecutive Patch Tuesday, since neither ...
- ConnectWise rotating code signing certificates due to security concerns
June 9, 2025
ConnectWise is updating the digital signing certificates used in ScreenConnect, ConnectWise Automate, and ConnectWise RMM due to concerns raised by a third-party researcher about how ScreenConnect handled certain configuration data in earlier versions. In addition to issuing new certificates, ConnectWise is releasing an update to improve how this configuration data is managed in ScreenConnect. This issue ...