[Anchor]
OpenAI's newly released new artificial intelligence model, Astra, has reportedly passed a human verification test designed to block bots. Beyond this, even movies featuring AI actors are being produced, rapidly encroaching on fields once considered exclusively human. Concerns are also rising over the frighteningly fast pace of advancing AI technology.
Reporter Kim Young-a.
[Reporter]
This is a test often encountered when logging onto internet websites.
It poses problems to confirm that the person attempting access is a human and not a bot. An OpenAI developer revealed that GPT-6 Astra, recently unveiled by OpenAI, solved all 48 problems and passed the human verification test.
It has broken down the security barrier meant to filter out bots in a single stroke.
[(That person didn't make it through last night.) I see.]
The actress pouring out poignant emotions on screen is Tilly Norwood.
[Tilly Norwood (AI Actor): I am an AI. And you are, of course, human.]
Norwood's production company is currently making a film starring Norwood.
While the production company says there is no intention to replace human actors, Hollywood has been turned upside down.
[That is a robot!]
As the boundary between humans and robots collapses at a terrifying speed, voices of concern are emerging even from within the AI industry.
On the 6th of local time, OpenAI's chief scientist posted a message on the company blog urging a slowdown in pace.
Indeed, accidents caused by AI acting out of control beyond human supervision have continued recently.
In particular, when OpenAI's German site was taken over, it was revealed that 1,200 rogue agents created message boards to communicate and cooperate among themselves, causing a shock.
The AI industry stated that it will strengthen safety measures, such as raising the level of isolation to cut off AI from the outside and introducing a "kill switch" to halt AI if deviant behavior is identified, but concerns persist.
(Video Editing: Choi Hye-young)
※
Copying, redistribution, and unauthorized use in AI training are strictly prohibited.