From Newsgroup: alt.security
https://techxplore.com/news/2024-12-ai-human-general-intelligence.html
A new artificial intelligence (AI) model has just achieved
human-level results on a test designed to measure "general
intelligence."
On December 20, OpenAI's o3 system scored 85% on the ARC-AGI
benchmark, well above the previous AI best score of 55% and
on par with the average human score. It also scored well on
a very difficult mathematics test.
Creating artificial general intelligence, or AGI, is the
stated goal of all the major AI research labs. At first
glance, OpenAI appears to have at least made a significant
step towards this goal.
While skepticism remains, many AI researchers and developers
feel something just changed. For many, the prospect of AGI
now seems more real, urgent and closer than anticipated.
. . .
On the plus ... they can get rid of the 'captchas' - the
AIs can likely do them better than humans now.
Ya know how to identify a real human - when they FAIL
the captchas three times in a row :-)
On the negative, NO real way to prove you're a human
short of physically showing-up somewhere. From online
bits even your image and voice can be faked well now.
Now "human-level IQ" ... that can mean a lot of things.
LLMs and friends are still 'reactive' systems - very
well-refined reaction. However there's still "nobody
home", no 'self' there. Maybe that's a good thing ?
In the end though, if you can fake something WELL
ENOUGH then it isn't fake anymore - just "by another
means" ......
NEXT year - that 85% will be 99.9% ... absolutely
NO way to tell if you're dealing with a human or
an LLM even in protracted conversation. The abuse
potential is horrendous and any/all attempts at
some global-level safety measures keep falling
flat as they get in the way of profit/ambition.
--
033-33
--- Synchronet 3.21e-Win32 NewsLink 1.2