Bloomberg reports data from an internal meeting at OpenAI that the company’s management team divides the progress of AI into 5 levels (Level 1-5), similar to how the ability levels of self-driving cars are classified by the Society of Automotive Engineers (SAE).
The AI levels at OpenAI are as follows:
Level 1 – Chatbots, AI with conversational language
Level 2 – Reasoners, human-level problem solving
Level 3 – Agents, systems that can take actions
Level 4 – Innovators, AI that can aid in invention
Level 5 – Organizations, AI that can do the work of an organization
The management team at OpenAI evaluates that the company’s technology is currently at Level 1 but approaching Level 2, which is a system capable of solving basic problems on its own, with abilities comparable to a human with a doctoral degree but without any tools. The meeting also discussed AI technology in reasoning.
OpenAI is not the first company to propose a 5-level hierarchy of AI capabilities, as Google DeepMind previously conducted research suggesting a similar classification. However, DeepMind believes that by 2023, only AI operating at Level 1 will exist.
Image depicting the levels of Google DeepMind
Source: Bloomberg
TLDR: OpenAI’s management categorized AI progress into 5 levels, with the company currently nearing Level 2, while Google DeepMind has also proposed a similar classification of AI capabilities.
Leave a Comment