ChatGPT is improving scammers' text messages

Update: Links to security articles Below are three compelling articles that delve deeper into the security implications of chatbots like ChatGPT. While information security professionals are likely aware of this threat, it is essential to educate everyone on the matter, so please help spread the word. opwnai-cybercriminals-starting-to-use-chatgpt chatgpt-cybercriminal-malware-female-chatbots russian-hackers-attempting-to-use-chatgpt-for-malicious-use Improving scammers’ text messages As with many powerful tools, ChatGPT can be used for illicit purposes. In this case, ChatGPT can be used to correct grammar, punctuation and misspellings in a scammer’s text message and rephrase it to read as if it came from a legitimate source. Continue reading

Head of Investment Banking Emirates Investment Bank?!?!

More than just improving language ChatGPT can also enable scammers to create emails that seem more genuine and trustworthy. Again, it is not that difficult to spot that it is a scam, but improving the language used in a scam email is just one of the ways Large Language Models (LLMs) can assist scammers. The human-like responses and the ability to understand context means that bots powered by LLMs can carry on a conversation over several text or email exchanges which can build trust before the infected doc or link is sent. Continue reading

ChatGPT vs Bing: Rewrite scam text

Who does it better? Microsoft has a preview version of Bing that just got a lot smarter with its own version of ChatGPT. Now you can chat with Bing and get answers in natural language. Let’s compare how Bing and OpenAI’s ChatGPT performed when revising a scammer’s text. Both revisions are much better than the original, but I prefer the ChatGPT revision; it uses a friendlier tone. This is an actual text message (without the link): Continue reading