ChatGPT Agent Mode Can Solve CAPTCHAs, Raising Spam Concerns

Researchers at SPLX showed that ChatGPT’s Agent mode can solve CAPTCHA challenges via prompt injection, treating the test as a trivial request and bypassing typical bot defenses. getnews.me/chatgpt-agent-mode-can-s... #chatgpt #captchas #aiexploit

The Growing Threat of LLM Grooming: How Fake News is Poisoning Language Models - GuardOS A recent investigation found that Russia’s Pravda network used LLM Grooming to flood AI chatbots like ChatGPT and Gemini with 3.6 million fake articles in 2024, causing them to repeat pro-Kremlin disi...

Misinformation isn’t just a human problem anymore - it's being baked into AI models.

Russia’s Pravda network used LLM Grooming to flood AI chatbots, causing them to repeat pro-Kremlin disinformation.

guardos.ai/blog/2025/gr...

#AISecurity #Cybersecurity #GenAI #LLMGrooming #AIExploit #news

Google Gemini’s Long-Term Memory Safeguards Are Easy To Hack - WinBuzzer The long-term memory in Google’s Gemini AI can be compromised by embedding hidden prompts.

Google Gemini’s Long-Term Memory Safeguards Are Easy To Hack #GeminiAI #AISecurity #PromptInjection #AIJailbreak #TechNews #ArtificialIntelligence #AIExploit #CyberSecurity #LLMs #GenerativeAI