OpenAI, the creator of ChatGPT, has introduced a 5-level test to assess how closely its AI model mimics human intelligence. This framework classifies AI capabilities into different tiers, starting from the basic conversational level to the most advanced organizational level.
According to reports, OpenAI shared this classification system during an internal meeting, revealing that they believe they are currently at the conversational understanding level and are close to achieving the reasoning level, where AI can solve problems akin to a human with a PhD.
The tiers are as follows:
– Level 1: Conversational Understanding
– Level 2: Reasoners
– Level 3: Agents
– Level 4: Innovators
– Level 5: Organisations
OpenAI executives demonstrated the capabilities of their GPT-4 AI model, showcasing human-like reasoning skills. The company emphasized their commitment to continuously testing new AI capabilities internally and staying at the forefront of technological advancements.
The decision to share these tiers with investors and the public underscores OpenAI’s transparency in monitoring the progress of AI development. As they strive to reach higher levels of AI intelligence, the company remains dedicated to ensuring a safe and responsible integration of AI technology into society.
In conclusion, OpenAI’s innovative approach to AI testing sets a new standard in the industry, highlighting the company’s efforts to push the boundaries of artificial intelligence while prioritizing ethical considerations.
