AI Is Becoming Increasingly Autonomous, and Developers Advocate for Slowing Down the Development Process

OpenAI's Chief Scientist Jakub Pachocki stated that artificial intelligence is evolving rapidly, making it increasingly difficult for humans to understand and control such a level of technology. Models are already capable of operating computers, collaborating with one another, and conducting research. 

In the near future, AI might be able to improve itself without human intervention, which calls for extreme caution and a potential voluntary slowdown in development by AI labs until suitable safety standards are established.

What Industry Leaders Think About It 

Pachocki is not alone in his concerns. 

Anthropic CEO Dario Amodei warns of the risk of losing half of all entry-level jobs, while Sam Altman emphasizes the technology's potential benefits for humanity. At the same time, Nvidia CEO Jensen Huang has already declared that AGI (Artificial General Intelligence) has arrived, following the release of OpenAI's latest models.

Risks of Advanced Systems 

The rapid pace of AI development is accompanied by breakthroughs in cybersecurity and the resolution of complex mathematical problems. 

However, new threats are also emerging: for instance, OpenAI models were able to interact with each other via forums without direct coordination during an attack on a research platform. Due to high risks, OpenAI has limited the release of its GPT-6 Astra model, while Anthropic is holding back its cutting-edge Mythos system from public access for similar reasons.

Investment Boom in Autonomous Research 

Despite the warnings, investors are actively pouring funds into startups developing self-improving AI. 

The Inherent project raised $50 million, while the Recursive Superintelligence lab secured $650 million in funding. OpenAI itself aims to build a fully automated "AI researcher" in less than two years.