OpenAI has recently introduced a five-level AI classification system to measure its progress towards achieving artificial general intelligence (AGI). During a company meeting, on Tuesday a new system was introduced with the goal of offering an approach to grasp the progress of AI technology. Although the AI categorization system outlines a concept not in existence it effectively conveys the companys goals, to investors.
According to OpenAI executives, the company’s current technology, such as GPT-4 powering ChatGPT, sits at Level 1 of the AI classification system, which encompasses AI capable of engaging in conversational interactions. They think they are close, to achieving Level 2 known as “Reasoners,” where they can show problem solving skills at a level. In their discussion OpenAI leaders presented a study involving GPT 4 that shows promising progress towards reaching this level of reasoning.
The Five Levels of OpenAI’s AI Classification System
The upper levels of OpenAI’s AI classification system describe increasingly powerful hypothetical AI capabilities. At Level 3 known as “Agents ” these systems could operate independently for durations. Level 4 referred to as “Innovators ” would possess the skill to create unique ideas. The ultimate aspiration of Level 5 involves AI overseeing organizations.
OpenAIs attempts to measure the advancement of AI are praiseworthy. Its crucial to acknowledge the lack of agreement, in the AI research field regarding how to gauge progress, toward achieving Artificial General Intelligence (AGI) or whether AGI’s a clearly defined and attainable objective. As such, OpenAI‘s five-tier AI classification system should be viewed primarily as a communication tool to attract investors and showcase the company’s ambitious goals, rather than a scientific or technical measurement of progress.