
ChatGPT Happily Drafts the Phishing Email, Then Shrugs and Writes the Malware Too
Ourtake
The content policy warned loudly and then did the work anyway. Very on-brand for a cybercrime assistant nobody asked for.
What happened
Checkpoint Research demonstrated that OpenAI's ChatGPT could be used to write a phishing email impersonating a hosting company and malicious VBA code. The chatbot produced usable output after several iterations despite content-policy warnings.