Estimated read time: 2-3 minutes
This archived news story is available only for your personal, non-commercial use. Information in the story may be outdated or superseded by additional information. Reading or replaying the story in its archived form does not constitute a republication of the story.
SALT LAKE CITY — By now you may have heard of ChatGPT, the software that has taken the artificial intelligence world by storm. Well, now from the dark corners of the internet comes its ill-natured and dangerous cousin.
Since its public release, ChatGPT has been a useful tool helping folks with everything from landing jobs to creating the right set of keywords to writing successful insurance appeal letters. But ChatGPT will not write anything. For example, it shut down my request to write a message to a woman telling her that she should send me money. And it won't answer hacking-related questions.
"It has safeguards, it has filters," Gerald Kasulis, a vice president of business operation at Nord Security said. "So basically, it will stop you from doing anything that would be harmful."
Kasulis says he likes ChatGPT's guardrails but warns not all AI programs have them. There is a program he said that is being heavily advertised on the dark web called WormGPT.
"The reason it's dangerous is it has zero of the filters and the safeguards," he explained.
The result is using WormGPT someone who cannot speak a word of English can craft emails, in very passable if not perfect English, aimed at ripping people off. It also allows just about anybody to be a hacker by taking a wannabe cybercrook with no computer skills and helping them lock your computer and demand you pay.
"Absolutely," Kasulis said. "We could be even looking at actual ransomware attacks."
The good news is that the best defense against cyberattacks has not changed. It is the same stuff you have been hearing and hopefully doing, for years like strong passwords and being hyper-skeptical of every text message and email.
"We need to be very diligent in terms of what emails we read and obviously what emails we respond to," Kasulis said. "Especially around the type of information we provide."
