OpenAI has reportedly created a new taxonomy that shows different levels of artificial intelligence (AI) development towards achieving super intelligent AI. According to the report, the classification system that OpenAI shared with its employees has five levels: The first level includes existing generative AI-powered conversational chatbots and the fifth will include the not-yet-achieved artificial general intelligence (AGI), an AI system that is considered powerful enough to take over the work of an entire organisation. The AI firm is also said to have shared details of a new project that uses the GPT-4 model and exhibits human-like reasoning capabilities – a capability that could take the model to another level.
OpenAI reportedly shares five levels of AGI
Citing an OpenAI spokesperson, a Bloomberg report said that the company shared the new classification system with its employees in an “all-hands meeting”. The company is also reportedly planning to share the classification with its investors and people outside the organization, although no timeline was shared for this.
At the moment, OpenAI is said to be at level one, however, citing people familiar with the matter, the report claims that the AI firm has shared new capabilities of GPT-4 that could be equivalent to level two. The most notable feature is reportedly the reasoning capability of the AI model. No other features were mentioned.
The publication also outlines five levels of AGI. These are:
- Level 1 — Chatbot: AI with conversational language
- Level 2 — Reasoner: Human-level problem solving
- Level 3 – Agents: systems that can perform actions
- Level 4 — Innovator: AI that can aid in invention
- Level 5 — Organization: AI that can do the work of an organization
Based on the reported classification, level one is where most AI systems are today with conversational chatbot capabilities. OpenAI is said to have reached level two with GPT-4 called Reasoner, however, this capability is yet to be demonstrated by the company.
The third level involves agents, which can take actions on behalf of people. This is probably the stage where AI systems can take certain actions independently based on behavioral patterns. This level also includes the integration of AI into real-world operations. According to the classification, the fourth level will give AI a certain level of innovation and it will be able to assist in inventing new technologies and products. Notably, true innovation is considered one of the defining traits of human intelligence.
Finally, level five or AGI is when an AI system can take over the functions of an entire organisation. This will far exceed the capability of humans and take AI to a super intelligent level.
