OpenAI has launched a ChatGPT Mac app lower than two weeks in the past – and there is already a safety challenge.
It seems that the ChatGPT app for Mac was storing customers’ conversations with the AI chatbot in plain textual content.
OpenAI has since fastened the problem with an replace, so it is necessary to obtain the most recent model of ChatGPT for Mac from the OpenAI web site.
ChatGPT for Mac safety vulnerability
Final week, Threads person @pvieito found that the lately launched ChatGPT app for Mac saved customers’ conversations with the AI chatbot in plain textual content in an unencrypted location.
Velocity of Mild Mashable
These conversations have been saved domestically, so a malicious actor would have needed to achieve entry to the person’s laptop to make use of these conversations for malicious functions. Nonetheless, as Mashable and different media shops lately reported, because of Microsoft’s points with its Callback operate, Dangerous actors are more and more getting access to unsuspecting victims’ computer systems through malicious distant strategies.
As @pvieito identified, OpenAI distributes its ChatGPT Mac app from its personal web site, not by Apple’s Mac App Retailer. Apple is thought for its strict pointers for builders who select to distribute apps by their official app shops. Nonetheless, third events who distribute by their very own platforms usually are not held to the identical safety requirements.
“We’re conscious of this challenge and have delivered a brand new model of the app that encrypts these conversations,” an OpenAI spokesperson mentioned. The sting in a press release final week. “We’re dedicated to offering a helpful person expertise whereas sustaining our excessive safety requirements as our expertise evolves.”
Whereas the problem has since been fastened, it could not have come at a worse time for OpenAI. The creator of ChatGPT and Apple simply introduced a Partnership final month and Apple is thought for prioritizing person privateness. The truth is, Apple’s incapacity to belief Meta concerning privateness points reportedly led the iPhone maker to desert its plans to make use of Meta’s AI fashions for its Apple Info ambitions.
OpenAI will must be extra proactive on privateness and safety points sooner or later in order that its credibility shouldn’t be questioned by its companions.