Artificial general intelligence (AGI) is the next evolution of artificial intelligence (AI) as it will have higher capabilities than humans. Efforts towards AGI have already been started by DeepMind and OpenAI. This week OpenAI lists the five levels AI needs to reach before it reaches AGI level.
Level 1 is a chatbot like ChatGPT. Level 2 is a system that can solve basic problems at a level comparable to a PhD holder. Level 3 is an AI agent capable of performing tasks on behalf of the user. Level 4 is AI that is capable of creating new innovations that don't exist yet. Level 5 which is the highest level is AI that is capable of doing the entire task of a human organization.
OpenAI says they are getting closer to Level 2. When Level 5 will be reached is still unknown. Sam Altman, CEO of OpenAI last year predicted that AGI could be achieved within five years.
Define the level of AI defined by OpenAI is very similar to that proposed by DeepMind last year. The five AGI level classifications proposed by DeepMind are emerging, competent, expert, virtuoso and superhuman. DeepMind also agreed that it is still in Level 1 through the use of chatbots.