IBM finds that ChatGPT can generate phishing emails nearly as convincing as a human

With just five simple prompts, ChatGPT built a convincing phishing email in minutes that got nearly as many clicks as a human-generated one.

ai chatgpt clicks computer security computers & electronics conversational-ai data security and privacy email emails generated generative ai human ibm large-language-models ml and deep learning network security and privacy nlp phishing phishing emails prompts security simple vb daily newsletter