OpenAI Unveils Roadmap for General Artificial Intelligence Development
OpenAI recently held an internal meeting where they discussed a new roadmap for developing Artificial General Intelligence (AGI), according to Bloomberg.
The company has broken down progress in AGI development into five levels:
- Level 1: Chatbots (AI with conversational language capabilities)
- Level 2: Reasoning AI (ability to solve problems at a human level)
- Level 3: Agents (systems capable of performing actions on behalf of a user)
- Level 4: Innovators (AI to assist in inventions and discoveries)
- Level 5: Organizations (AI capable of performing the work of an entire company)
Sources indicate that OpenAI is currently at Level 1. They claim Level 2 will be able to solve problems at the level of a PhD, with more detailed information about the other levels to be released later.
The meeting also included a presentation of a project using their GPT-4 model, aimed at testing skills related to human thinking. Finally, OpenAI co-founder Ilya Sutskever announced the launch of a new company, Safe Superintelligence Inc. (SSI), dedicated to developing safe superintelligence.
"We will strive to create safe superintelligence, focused on a single task, with a single goal and a single product. We will achieve this through revolutionary breakthroughs made by a small team of experienced developers. Join us," Sutskever stated.
