OpenAI has introduced a five-tier framework to monitor its advancement toward developing artificial intelligence that can rival and even surpass human capabilities.

The initiative is the latest in the startup’s efforts to enhance public understanding of AI safety and was shared with staff during a company-wide meeting on Tuesday, July 9. OpenAI intends to present the levels to investors and other stakeholders, which span from conversational AI (Level 1) to AI that can independently operate an entire organization (Level 5).

During the meeting, OpenAI executives informed employees that the company is currently at the first level but is nearing the second level, known as Reasoners. This tier represents AI systems capable of basic problem-solving tasks comparable to a human with a doctorate-level education.

In the same session, OpenAI’s leadership demonstrated a research project featuring the GPT-4 AI model, showcasing new skills indicative of human-like reasoning. For years, the company has been working towards achieving what is often referred to as artificial general intelligence (AGI), which entails creating computers that can outperform humans in most tasks. Such systems do not yet exist, though OpenAI CEO Sam Altman has previously suggested that AGI might be achievable later this decade.

Also Read: The Most AI-Proof Career Opportunities In The Middle East

Determining the criteria for AGI has been a topic of ongoing debate among AI researchers. In a paper published in November 2023, researchers at Google DeepMind proposed a framework of five ascending AI levels, including “expert” and “superhuman”, which resembles the classification system used in the automotive industry for self-driving cars.

According to OpenAI’s proposed levels, the third tier on the road to AGI is called Agents, representing AI systems that can perform tasks autonomously over several days. Level 4 describes AI that can generate new innovations, while the highest level, Organizations, refers to AI capable of managing entire enterprises.

The framework, developed by OpenAI executives and senior leaders, is considered a work in progress. The company plans to collect feedback from employees, investors, and its board, with the possibility of refining the levels over time.