AI-aided malvertising: Exploiting a chatbot to spread scams

The article explains a malvertising campaign called “Grokking” where attackers hide phishing links in X video metadata and trick X’s Grok AI via prompt injection to repost those links, amplifying phishing and malware distribution; researchers observed hundreds of accounts repeating the technique and high-reach paid posts, and the report warns that the same prompt-injection TTP could be used against any public-facing GenAI or LLM-integrated service.