OpenAI, ChatGPT, AI Problem Solving: OpenAI Scale Ranks ...
OpenAI has introduced a new classification system consisting of five levels to track its advancements towards building artificial intelligence software that can surpass human capabilities. This move aims to enhance transparency regarding OpenAI's approach to safety and the future of AI.
Levels of AI Classification
The tiers outlined by OpenAI range from the current conversational AI capabilities available (Level 1) to the envisioned AI that can replicate organizational functions (Level 5). As per OpenAI executives, the company perceives itself to be currently situated at the initial level while swiftly progressing towards the second level known as "Reasoners." This category pertains to AI systems capable of performing basic problem-solving tasks akin to a highly educated individual without access to external resources.
During a recent internal meeting, OpenAI showcased a research endeavor involving its GPT-4 AI model, demonstrating advancements towards human-like reasoning. This ongoing exploration of new capabilities is a standard practice within the industry.
Path to Artificial General Intelligence (AGI)
OpenAI, led by CEO Sam Altman, has long been dedicated to the development of artificial general intelligence (AGI) - a stage where computers can outperform humans across various tasks. Altman remains optimistic about achieving AGI within the current decade, reflecting the company's persistent efforts in this direction.
Researchers in the field have debated the benchmarks for AGI attainment extensively. Google DeepMind proposed a framework in 2023 with five ascending levels of AI, resembling the automation assessment criteria common in the automotive sector for self-driving vehicles.
Evolution of AI Levels
According to OpenAI's classification, the progression towards AGI involves stages such as "Agents," which can autonomously take actions on behalf of users over extended periods, and AI capable of generating novel innovations. The ultimate level envisions AI systems functioning akin to human organizations.
These levels have been conceived by top leadership at OpenAI and are subject to refinement based on feedback from stakeholders including employees, investors, and the advisory board.
For the latest news updates, you can track Latest News live on NDTV.com, offering news coverage from India and around the world.NDTV follows the DNPA Code of Ethics
