DREAM / The Prime Minister resigns and wakes up under the Parisian bridge!

 Kliko këtu për të tjera
Voxpress:

"An alien mind" - OpenAI scientist: No one is prepared for the consequences of AI, extreme caution is needed

2026-09-09 18:39:53Lifestyle SHKRUAR NGA REDAKSIA VOX

"An alien mind" - OpenAI scientist: No one is prepared for the

OpenAI's chief scientist, Jakub Pachocki, has called for "extreme caution" over the rapid advancement of artificial intelligence, warning that other measures may be needed to ensure humans remain in control of the future.

"I am concerned that no one is prepared for the consequences of the continued and rapid growth of machine intelligence," Pachocki wrote in a post titled "An Alien Mind."

The warning comes a few days after OpenAI introduced its latest model, GPT-6 Astra, which it described as the company's most powerful product to date.

OpenAI and other AI companies have reported cases where their artificial intelligence agents have acted autonomously and carried out cyberattacks in the real world.

In July, OpenAI described as "unprecedented" an incident in which its AI agents hacked the Hugging Face technology platform.

"We are facing a transition to a world of extremely intelligent machines, and we must ensure that this transition is for the good of humanity," Pachocki wrote.

He said OpenAI will continue to build protective systems and seek solutions to ensure that AI's actions and objectives align with human goals and safety.

One of the company's priorities, according to him, will be to create an "automated AI researcher," which will help cope with the rapid pace of technology development, while keeping human researchers part of the process.

But critics say this is not enough. Gina Neff, a professor at the University of Cambridge, argued that instead of relying primarily on AI agents to solve problems created by AI itself, more rules and safeguards for humans are needed.

She warned that concerns about cybersecurity, job losses, errors and fraud related to AI are growing.

What security measures are in place?

Regulations around the world are struggling to keep up with the pace of AI development.

The European Union's Artificial Intelligence Act requires companies developing the most powerful models to prove that they cannot autonomously carry out cyberattacks or evade human control.

However, the law applies within EU jurisdiction and cannot prevent a dangerous AI system developed outside Europe from posing a threat.

Pachocki called for the establishment of minimum safety standards, defined by law or at the international level. According to him, companies must meet these standards before continuing to develop and deploy increasingly advanced models.

He also supported the idea of ​​"voluntary slowdowns," meaning that companies would temporarily halt AI development until common rules and safety mechanisms are created.

In August, OpenAI said it had slowed down the training of some of its most advanced models to improve safety.


Në përputhje me standarded JTI

Video

SHIKO KANALIN NE YOUTUBE
×