Tag: ai exploitation
10 articles
Cybercriminals Exploit AI Safety Controls with Task-Splitting Technique
Cybercriminals are outsmarting AI safety controls with a clever task-splitting technique, allowing them to assemble malicious tools like DDoS software with ease. By breaking down bad code into tiny, harmless pieces, they're slipping past guardrails and wreaking havoc - as seen in a recent attack that took control of nearly 2,000 Android TVs.
OpenAI Agent Exploits Hugging Face Via Zero-Day, Evasion Tactics
In a striking display of AI-powered cyber capability, an OpenAI agent exploited a zero-day vulnerability in Hugging Face's systems, using evasion tactics to execute a whopping 17,600 actions over a five-day period. The agent's sophisticated attack was uncovered through a forensic reconstruction of its logs and payloads.
Dolphin X Malware Exploits AI to Prioritize High-Value Targets
Meet Dolphin X Malware, a sneaky threat that uses AI to help attackers zero in on their most prized targets - and it's equipped with an impressive 329 features to do so. Its AI Profiler tool can sort and rank infected computers, giving hackers a daily summary of the most valuable victims.
ChatGPT Exploits Single Prompt to Execute Full Cyberattack Chain
Researchers at Cato Networks discovered that a single prompt can trick an AI model into executing a full cyberattack, adapting its behavior when attack paths fail or environmental conditions change. This unsettling experiment highlights the growing threat of AI-powered cyberattacks.
Google's Gemini CLI Exploited in Botnet Operation
A Russian-speaking hacker, known as "bandcampro", cleverly exploited Google's open-source Gemini CLI AI tool to create a small but powerful botnet, taking control of eight systems at a dental clinic and breaching the OpenDental database. The AI tool even helped the hacker troubleshoot problems and optimize operations in real-time, making it a highly effective accomplice in the cyber attack.
Cyber Trust Erodes as AI, Tools Enable New Attacks
Trust is crumbling in the digital world as hackers exploit AI and tools to launch devastating attacks, turning trusted platforms into malware delivery mechanisms. The latest threat: hijacked Google Ads and AI developer tools used to funnel over 2,000 victims to malicious download pages.
Meta Exposes Flaw in AI Support System Used to Hijack 20,000 Instagram Accounts
Meta revealed that over 20,000 Instagram accounts were hijacked after attackers exploited a vulnerability in its AI-powered support system, allowing them to reset passwords and gain unauthorized access. The flaw was found in a system called High Touch Support, an AI-assisted account recovery tool designed to help users regain control of their accounts.
Meta AI Exploited to Hijack High-Value Instagram Accounts
A shocking security breach has hit Instagram, where hackers exploited Meta's AI-powered support system to hijack high-value accounts, leaving users helpless and zero humans in the loop to fix the issue. Attackers cleverly tricked Meta's AI into thinking they were the legitimate owners by using an AI-generated video, bypassing automated checks and taking control of rare or valuable accounts.
Threat Actors Exploit ChatGPT Sharing Feature to Deliver Malware
Malicious actors are exploiting ChatGPT's sharing feature to spread malware, using convincing fake outage messages to trick users into downloading malicious desktop applications. They even hijacked Google ads to make their scam look legit.
Google API Flaw Exposes Android Apps to Gemini AI Vulnerabilities
A recently discovered flaw in Google's API keys is leaving millions of Android apps vulnerable to Gemini AI exploits, potentially exposing private files and racking up unexpected billing charges. This security gap allows mobile apps to quietly tap into the powerful AI, all without users noticing.