Fb proprietor Meta stated on Wednesday it had uncovered malware purveyors leveraging public interest in ChatGPT to lure customers into downloading malicious apps and browser extensions, likening the phenomenon to cryptocurrency scams.
Since March, the social media big has discovered round 10 malware households and greater than 1,000 malicious hyperlinks that had been promoted as instruments that includes the favored synthetic intelligence-powered chatbot, it said in a report.
In some circumstances, the malware delivered working ChatGPT performance alongside abusive recordsdata, the corporate stated.
Talking at a press briefing on the report, Meta Chief Data Safety Officer Man Rosen stated that for dangerous actors, “ChatGPT is the brand new crypto.”
Rosen and different Meta executives stated the corporate was getting ready its defenses for a wide range of potential abuses linked to generative AI applied sciences like ChatGPT, which might shortly create human-like writing, music and art.
Lawmakers have flagged the tools as more likely to make on-line disinformation campaigns simpler to propagate.
Requested if generative AI was already being utilized in data operations, the executives stated it was nonetheless early, although Rosen stated he anticipated “dangerous actors” to make use of the applied sciences to “attempt to velocity up and maybe scale up” their actions.
