ChatGPT maker OpenAI now has a scale to rank its AI

ChatGPT maker OpenAI now has a scale to rank its AI

An OpenAI spokesperson has shared the company’s new 5-tier system for ranking its progress toward achieving Artificial General Intelligence (AGI), reports Bloomberg.

The levels, which were announced by the company behind ChatGPT internally at an all-hands meeting before being shared externally, are designed to guide thinking about artificial intelligence (AI) and its capabilities as the company works to develop models with real reasoning capabilities.

Levels of AI on the path to artificial general intelligence

The levels in the system were outlined like this:

  • Tier 1 – Chatbots, AI with conversational language. This is where OpenAI believes their current GPT4 models are at currently

  • Tier 2 – Reasoners, human-level problem solving (described as having the reasoning capabilities of a person with a PhD but without any tools)

  • Tier 3 – Agents, systems that can take actions on a user’s behalf.

  • Tier 4 – Innovators, AI that can come up with its own innovations

  • Tier 5 – Organizations, AI that can do the work of an entire organization

AGI is the long-term goal for many companies involved in the AI arms race, including Mark Zuckerberg’s Meta.

While OpenAI believes they are currently at level 1, their spokesperson said they are on the cusp of reaching the second level, “Reasoners”.

During the all-hands meeting where the new levels were announced, OpenAI also demonstrated some new research centered around its GPT4 model, which they believe shows skills approaching human-level reasoning.

The levels, which were designed by OpenAI’s senior leadership team and executives, are not considered final. As the organization gathers feedback and additional input from its employees and investors, it may alter the levels and definitions over time to better fit the broader understanding of AI progress.

OpenAI’s stated mission is to develop “safe and beneficial artificial general intelligence for the benefit of humanity”, however, earlier this year the company effectively dissolved its safety-oriented Superalignment group after the departure of Chief Scientist and co-founder Ilya Sutskever. This has led to questions being raised about whether the company can truly live up to its mission statement.

Featured image credit: generated with Ideogram