OpenAI, the company behind the popular language model ChatGPT, has unveiled a new 5-tier system to assess its advancements towards achieving Artificial General Intelligence (AGI). This development was shared by an OpenAI spokesperson during an all-hands meeting and subsequently made public.
The tiered system is designed to provide a framework for understanding the capabilities of AI models as OpenAI continues its efforts to develop models with genuine reasoning abilities. The levels outlined in the system are as follows:
Tier 1 – Chatbots, AI with conversational language: OpenAI believes that their current GPT4 models fall under this category, representing AI systems capable of engaging in conversational interactions.
Tier 2 – Reasoners, human-level problem solving: This tier encompasses AI models that possess problem-solving capabilities comparable to those of humans. The spokesperson did not provide further details on this level.
Tier 3 – Innovators, surpassing human-level performance: The third tier represents AI systems that surpass human-level performance in specific domains. However, the spokesperson did not elaborate on the specific domains or provide examples.
Tier 4 – Societally beneficial AI: This tier focuses on AI models that are designed to benefit society as a whole, taking into account ethical considerations and potential risks. The spokesperson emphasized the importance of aligning AI systems with human values.
Tier 5 – AGI, surpassing human-level performance in all domains: The final tier represents the ultimate goal of OpenAI’s research – the development of Artificial General Intelligence. AGI refers to AI systems that can outperform humans across a wide range of tasks and exhibit a high level of adaptability.
OpenAI’s spokesperson, Ali Rees, highlighted that the tiered system aims to guide discussions and foster a better understanding of AI’s progress towards AGI. However, no specific timeline or roadmap for achieving AGI was provided.
It is important to note that OpenAI’s announcement does not indicate the immediate availability of AI models at each tier. Rather, it serves as a framework to categorize and evaluate the company’s progress in developing increasingly advanced AI systems.