On 1/24/25 7:48 AM, glen wrote:
https://abnormalsecurity.com/blog/ghostgpt-uncensored-ai-chatbot?ref=metacurity.com
"GhostGPT is a chatbot specifically designed to cater to
cybercriminals. It likely uses a wrapper to connect to a jailbroken
version of ChatGPT or an open-source large language model (LLM),
effectively removing any ethical safeguards. By eliminating the
ethical and safety restrictions typically built into AI models,
GhostGPT can provide direct, unfiltered answers to sensitive or
harmful queries that would be blocked or flagged by traditional AI
systems."
oddly, considering chatting with such an LLM, feels like choosing to
engage with a sociopath? Or maybe a psychopath? Why I find Trump's
fascination with Hannibal Lector effing creepy. I *have* pushed GPT up
to some of it's limits and even enjoyed doing so, but not sure how I'd
feel about it not having any limits?
Seems like a good design for a "honeypot" trap... ATF or FBI creating
an LLM happy to collude with would-be terrorists on all kinds of things
that perhaps legal or at least ethical boundaries would not allow field
agents to do? Might sweep up every highly disturbed teen with
ideation and access to school shooting tools before they actually went
into the schools and started shooting? Or lead to a whole new suite of
Trump-Pardons for what he would call "hostages"?
Shades of PK Dick's "Minority Report"?
.- .-.. .-.. / ..-. --- --- - . .-. ... / .- .-. . / .-- .-. --- -. --. / ...
--- -- . / .- .-. . / ..- ... . ..-. ..- .-..
FRIAM Applied Complexity Group listserv
Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom
https://bit.ly/virtualfriam
to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com
FRIAM-COMIC http://friam-comic.blogspot.com/
archives: 5/2017 thru present https://redfish.com/pipermail/friam_redfish.com/
1/2003 thru 6/2021 http://friam.383.s1.nabble.com/
