OpenAI, a leading player in the development of artificial intelligence (AI) systems, has introduced a new classification system to track its progress towards building AI software capable of surpassing human intelligence.
The company unveiled this system during an all-employees gathering, to provide greater transparency and understanding of its approach to safety and the future of AI.
The system consists of five levels, each representing a different stage of AI development. At the first level, "Chatbot", the AI is capable of engaging in conversational language with humans. As the levels progress, the AI becomes more advanced. The ultimate level is where the AI can execute tasks as competent as an entire organization.
Currently, OpenAI believes that it is on the first level and is approaching the second level, which it refers to as "Reasoners."This level involves AI systems that can perform basic problem-solving tasks at a level comparable to a human educated at a doctorate level without any external source of information.
Progress Levels Towards AGI
The introduction of this classification system by OpenAI aims to provide a structured framework for tracking the advancements and setting benchmarks in the pursuit of artificial general intelligence (AGI).
AGI refers to AI systems that possess human-like intelligence and surpass humans in economically valuable tasks. It is considered the broad goal for AI developers.
OpenAI's scale breaks down the progress towards AGI into five distinct levels. At the first level, we find systems like ChatGPT, which already demonstrate the ability to engage in conversations with humans.
At level 3, "Agents", the AI is capable of performing tasks on behalf of the user without their presence. At level 4, "Innovators", the AI can generate new ideas and concepts, showcasing its creativity. Finally, at level 5, "Organizations", the AI can not only perform tasks for individuals but can effectively take over tasks for entire organizations.
The introduction of this classification system is a significant step towards providing a comprehensive framework for evaluating AI models, not just for OpenAI but for the entire AI community. It allows for better tracking and understanding of AI's progress towards human-level intelligence.
However, experts have varying opinions on the timeline for achieving AGI. Some believe it could happen within the next five years, while others think it may take much longer.
The challenges associated with AGI development are substantial. They include the need for significant computing power, financial and technological hurdles, and ethical and safety concerns.
There is also legitimate concern about the societal implications of AGI. OpenAI's recent reorganization, which resulted in the dissolution of its safety team and the resignation of key researchers, raised concerns about the company's position on ethical and safety issues associated with AGI.
OpenAI has been constantly pushing the boundaries of AI development, with plans for further advancements in the pipeline.
One of these upcoming models is GPT-5, which OpenAI CEO Sam Altman has described as a "significant leap forward." While details about GPT-5 are still limited, Altman's remarks imply that it will address the limitations and errors seen in previous models.