Former Google CEO Eric Schmidt identified that in a single to 2 years, artificial intelligence (AI) will find a way to operate by itself, and that it’s time to contemplate introducing a ‘kill switch’ to stop AI operation. Likewise, there was a recent surge in expert opinions that artificial general intelligence (AGI) may emerge in a single to 2 years.
Former CEO Schmidt appeared on ABC News’ ‘This Week’ on the sixteenth and said that it’s time to noticeably consider introducing an AI kill switch.
“I’ve never seen innovation on this scale before,” he said on the printed. “Soon, computers will find a way to run on their very own and choose what to do on their very own.”
For this reason, “when the system starts to enhance itself, it might be time to drag the plug,” he said. “In theory, there needs to be someone to drag the plug.” In other words, the federal government should have a tool that physically blocks AI immediately, like a nuclear button.
The kill switch is a difficulty that has been discussed so much before. Last February, UC Berkeley professor Stuart Russell, a number one AI scholar, attended a domestic event and advocated for its introduction. Also, in May of last 12 months, major AI corporations comparable to OpenAI and Google announced that they might conform to introduce a sophisticated model kill switch through the ‘AI Summit’.
On the time, this was considered a future story of achieving AGI. Nevertheless, in recent months, claims that we’re on the verge of achieving AGI have increased significantly.
Sam Altman, CEO of OpenAI, is a representative example. Starting this summer, he emphasized that AGI is closer than expected and that the actual problem is superintelligence. This logic became stronger with the appearance of the inference model ‘o1’. Specifically, an Open AI worker claimed that “AGI has already been achieved.”
Specifically, Ilya Sutskever, former Open AI chief scientist, predicted on the ‘Newlips’ conference held last week, “If AI is provided with agent and reasoning capabilities at the identical time, it’s going to have deeper understanding and even self-awareness.” Former CEO Schmidt’s remarks are also consistent with this context.
One other person famous for such remarks is CEO Elon Musk. Last March, he predicted AGI’s appearance, saying, “Perhaps next 12 months, AI shall be smarter than any individual.” On the time, there was loads of ridicule, but now the atmosphere has modified.
Former CEO Schmidt explained that the world shouldn’t be yet ready for the impact of the emergence of AI. “It means everyone has a physician of their pocket,” he explained. “We now have no idea what it means to provide that power to each individual.”
The answer was to construct a separate system to screen out the worst-case scenario that AI would bring. “Humans won’t find a way to police AI,” he said, “but AI systems will find a way to police AI.”
Meanwhile, he made a series of remarks targeting the incoming Trump administration, saying that China has surprisingly caught up with america over the past six months.
Reporter Lim Da-jun ydj@aitimes.com
