Topic: data poisoning

October 10, 2025

Google Ignores Critical Gemini ASCII Attack

A newly discovered ASCII smuggling vulnerability in Google's Gemini AI allows attackers to use invisible Unicode characters to manipulate the system, potentially spreading false information or accessing unauthorized data. This security flaw affects multiple AI platforms including Google Gemini, D...

AI's Hidden Threat: Directed Bias Attacks on Brands

AI systems shape brand perception but lack truth discernment, potentially amplifying biased or false information and causing reputational harm. Malicious actors can intentionally poison AI data through tactics like directed bias attacks, semantic misdirection, and prompt manipulation to spread mi...

1 in 4 UK and US Firms Hit by Data Poisoning Attacks

One in four organizations in the UK and US have experienced data poisoning attacks, where malicious actors corrupt AI training data to disrupt operations or bypass security measures. Unauthorized employee use of generative AI tools, known as "shadow AI," poses risks such as data leaks and complia...