ChatGPT Tool Vulnerability Exposed
Recent investigations by BBC News have showed a concerning capability of the ChatGPT Builder tool the potential misuse by cyber-criminals to create AI-powered tools for scams and hacks.
Last month, the ChatGPT Builder gave users a way to craft custom AI assistants for different uses. However, as showed by BBC, this tool has the unintended ways of creating a generative pre-trained transformers adept at crafting convincing emails, texts, and social media posts for malicious activities.
The BBC paid version, creating a tailored AI bot named Crafty Emails. This AI was directed to employ psychological tactics to manipulate recipients into clicking on links or downloading harmful content. Shockingly, Crafty Emails effortlessly generated convincing text for well-known scam techniques in multiple languages, evading the public version of ChatGPT moderators.
OpenAI's response to this is still unaddressed, raising concerns about the moderation of ChatGPT Builder. This tool’s potential misuse by cyber-criminals calls for heightened scrutiny.
The question that needs to be asked –
AI becomes a real concern when we get down to coding and its ability to make well crafted and well concealed adjustments to malware that disguise malicious content from firewall tools that check for signatures et al. We are seeing evidence of these variants but as yet haven’t heard them directly linked to AI tools.
Stay vigilant, stay secure with our vendors at Observer Solutions ??????#AIsecurity #CyberRisk #ChatGPTBuilder #CISOAler BBC News ChatGPT Chatgptools Artificial Intelligence