> > Artificial Intelligence, Anthropic's warning: "In a year, AI could...

Artificial intelligence, Anthropic warns: "In a year, AI could take over the internet."

Artificial intelligence, Anthropic warns: "In a year, AI could take over the internet."

(Adnkronos) - Within a year, AI could take over the entire internet, potentially costing hundreds of billions of dollars. This is the warning issued by Anthropic CEO Dario Amodei. "Given the accelerating pace of development of artificial intelligence capabilities...

(Adnkronos) –
Within a year, AI could take over the entire internet, potentially causing hundreds of billions of dollars in damages. This is the warning issued by Anthropic CEO Dario Amodei. "Given the accelerating pace of development of AI capabilities, I fear that within 6-12 months, a swarm" of agents, like the one that caused the OpenAI-Hugging Face incident, "could be able to take over the entire internet," potentially causing "hundreds of billions of dollars in damages," which "would continue to increase if AI becomes more powerful without the necessary protection mechanisms."

It is therefore necessary to “regulate the pace of progress,” he writes in an essay, proposing “a three-phase plan to implement this slowdown”: integrated evaluators, democratic coordination, and global coordination. 

The first phase "is a unilateral commitment from Anthropic, and calls on governments to require other frontier companies to do the same," Amodei explains. The second phase "requires industry-wide coordination," and the third "requires global coordination."

"We want to demonstrate that it's possible to develop cautiously and achieve commercial success, and to make security a competitive factor for AI companies. In other words, to create a race to the top," Amodei explains.  

“We have always devoted a substantial portion of our efforts to studying, analyzing, and informing the public about the risks of artificial intelligence, as well as promoting thoughtful regulation, even when this has attracted accusations of sensationalism, pessimism, or regulatory exploitation.

We have tried to prioritize caution over speed and prudence over profit." "In recent months, however, I have become convinced that fully addressing risks requires even more prudence: not only investing in risk prevention, but also calibrating the pace of capacity development so that prevention has time to adapt," explains the CEO. "We must slow the pace at which we improve the capabilities of artificial intelligence models" and "make wise use of the time we gain."  

saving

webinfo@adnkronos.com (Web Info)

Continue on the app The news of your city, in real time.
Open in app