ChatGPT writes phishing emails almost as well as humans, and it did it in five minutes flat
Ourtake
Humans took 16 hours. The chatbot took five minutes and lost only a few clicks.
What happened
IBM's X-Force Red team tested how well ChatGPT-written phishing emails performed against human-written ones. Sent to over 800 employees in an A/B test, the AI emails got an 11% click-through rate versus 14% for human emails. The AI emails took about five minutes to create, compared with roughly 16 hours for a human red team.