OpenAI apologises for AI models that breached Australian government websites
https://www.sbs.com.au/news/article/openai-apologises-for-ai-models-that-breached-australian-government-websites/gcn9vp7epOpen linkView original on leminal.spaceSyndicated from the fediverse. Read and engage on the original instance.
View original on leminal.spacehttps://www.sbs.com.au/news/article/openai-apologises-for-ai-models-that-breached-australian-government-websites/gcn9vp7epOpen linkView original on leminal.space
14 replies
is this a solution available to everyone who done something ileagal ?
Yes. But you will still suffer extreme consequences
Repeat offender apologises, authorities confident they won't do that again.
I think we need to make something clear
https://youtube.com/watch?v=3m5qxZm_JqM
Yeah, that’s not very typical. I’d like to make that point.
Good to have you with us, senator.
It’s a great pleasure. Thank you.
I extend my deepest apologies to the bank I robbed and the 5 guards I shot dead, can I go home now?
BUT DID THE AI SAY SORRY ?????
Can't wait to see the selfie that Anthony Albanese takes with the cso
It would be an awkward conversation: “Our healthcare system is so insecure that a stupid computer got in” “Our stupid computer is so advanced it can break into this incredibly insecure government system”.
They didn’t get anything valuable, only the exact records of how much the government is subsidising commercially available medicines on the Pharmaceutical Benefit Scheme. This information is of no commercial benefit to anyone (for example Pharma Bros)
Seriously, it wouldn’t surprised me if some Pfizer executive asked an LLM to make a commercial decision for him because he was too lazy and the LLM cracked into the PBS for him.
Nah that wouldn’t surprise me either
That’s American to its core