GPT-J: Automatically generating malicious code and phishing emails
GPT-J: Automatically generating malicious code and phishing emails
GPT-J: Automatically generating malicious code and phishing emails.
The current giant in the field of artificial intelligence, OpenAI, has launched appeals more than once, saying that people need artificial intelligence research to avoid unnecessary problems caused by AI without ethical constraints.
News on July 17, recently, some hackers have shown what artificial intelligence can do in the case of “no moral restrictions”.
According to the information provided by the security analysis company SlashNext, some hackers have developed a generative artificial intelligence called “WormGPT” by using an open source AI model called GPT-J.
After testing by SlashNext, WormGPT can freely generate malicious code written in Python language, and can generate extremely misleading phishing emails.
And this kind of performance is only based on GPT-J, an open source model born in 2021, with an older architecture and poor performance.
It is conceivable that if the artificial intelligence of major manufacturers has no “moral” restrictions, what serious impact it will have.
Therefore, SlashNext believes that with the popularization of AI technology, such AI tools “completely born for cyber attacks” may be abused by hackers, posing a major threat to user security.
