Thieves can use ChatGPT to write convincing scam messages with human-like language, experts warn

ByJason Knowles and Ann Pistone WLS logo
Tuesday, March 14, 2023
Thieves can use ChatGPT to write convincing scam messages: experts
What is ChatGPT? Experts are warning that thieves can use OpenAI artificial intelligence to create convincing online scams.

CHICAGO (WLS) -- There's an I-Team alert on Tuesday afternoon about artificial intelligence, and how scammers can now use it to send you texts and emails that may look very real.

It's a legitimate and popular online tool called ChatGPT. This new artificial intelligence technology creates well-written texts for you, and can help draft letters. But, tech experts said thieves can also use it for wrongdoing, because it easily creates summaries or stories in human-like language.

"It's become hugely popular very quickly," said Alex Hamerstone with the tech security group TrustedSec.

SEE ALSO | Microsoft brings Bing chatbot to phones after curbing quirks

Like anything popular, Hamerstone said, criminals may find ways to abuse the technology.

"They're able to craft well-written scam emails," he said.

If a scammer uses ChatGPT, it makes the job easier, and could increase the volume of phishing attacks, which may convince you to click a malicious link or hand over personal information. Plus, ChatGPT summaries may now be very well-written with perfect grammar.

That means you want to be on guard now for texts and emails that look too well-written. That goes against the traditional warnings of looking out for poorly written texts and emails.

RELATED | Vanderbilt University apologizes after using ChatGPT to console students

"One of the biggest pieces of advice that we've always given over the years for these phishing attacks, or phishing emails and other things are to look for poorly written emails. You know, a lot of the people who write these are based overseas, and maybe English is in the first language, and you can oftentimes identify," Hamerstone said.

A spokesperson for OpenAI, the company that developed Chat GPT, said, "We don't want ChatGPT to be used for misleading purposes anywhere." They pointed the I-Team to the company's other tech, called Classifier tool, "which helps differentiate between AI-written and human-written text."

"With these new AI technologies that help people write, it's almost impossible to discern some of these fishing attacks from a legitimate email at this point," Hamerstone said.

READ MORE | New Bing AI chatbot raises alarms, with users reporting insults, threats, comparisons to Hitler

One way to avoid phishing texts or emails all together is to delete any and all unsolicited correspondence. If you think a company or government agency legitimately reached out to you, call them directly to verify.

You can report phishing emails and texts to the Federal Trade Commission.