SCROLL

August 2024

OpenAI Charts the Path: Is Humanity Ready to Follow?

OpenAI Charts the Path: Is Humanity Ready to Follow?

In an effort to provide a clearer and more comprehensible view of its progress in building artificial intelligence (AI), OpenAI has developed a new level system that classifies advancements toward creating AI software capable of surpassing humans in various tasks. This system, recently shared with employees and planned to be presented to investors and the general public, aims to offer a more detailed perspective on the evolution of AI and the company's expectations.

 

 

A Level System to Understand Progress

OpenAI's Level System

  • Level 1: Current conversational AI, capable of interacting in natural language.
  • Level 2 "Reasoners": Systems that perform problem-solving tasks comparable to those of a person with a Ph.D., without using external tools. OpenAI is close to achieving this level.
  • Level 3 "Agents": AI that can manage tasks over several days on behalf of a user.
  • Level 4: Systems that can innovate and create new solutions.
  • Level 5: AI that can perform the complete work of an organization, representing the highest level of advancement.

 

Advancements Toward Human-Like Reasoning

OpenAI aims to develop Artificial General Intelligence (AGI), an AI that surpasses humans in most tasks. Although AGI systems do not yet exist, OpenAI is confident in achieving this goal within this decade. The recent demonstration of the GPT-4 model highlights significant progress toward more human-like reasoning, indicating that the company is advancing on its path to AGI.

 

 

What Do the Geeks Think?

The tech community is closely watching the evolution of OpenAI's level system. Most believe that this approach provides a clear structure for understanding AI advancements. The level classification allows for a more precise evaluation and effectively communicates the company’s achievements and expectations.

However, some challenges are recognized. Implementing AI at advanced levels, such as Agents and Organizations, raises questions about performance and integration into real systems. Additionally, accessibility and ethics in AI development are critical issues that must be considered as we advance toward more sophisticated levels.

 

Stay up to date with the latest digital trends.

Subscribe to our newsletter!