Artificial General Intelligence (AGI) is a concept that has fascinated scientists, researchers, and technologists for years. AGI refers to AI that can perform any intellectual task that a human being can. OpenAI, a leading organization in AI research, has recently introduced a five-level framework to track its progress toward this goal. This blog will delve into these levels and explore their implications.

At this level, AI systems can hold conversations with humans. This is the stage where most current AI, including OpenAI’s ChatGPT and other chatbots like Claude, reside. These systems can understand and respond to human language, making them useful for customer service, virtual assistance, and other applications requiring human-like interaction.

Level 2 represents AI systems that can solve problems at a human level. OpenAI is nearing the development of these ‘Reasoners.’ Current models like ChatGPT occasionally demonstrate reasoning abilities, but achieving expert-level reasoning is the next milestone. This advancement is significant because it addresses the issue of AI hallucinations—where AI generates convincing but incorrect responses. A Level 2 AI would provide more accurate responses and evaluate the reliability of its answers, thus overcoming the limitations seen in enterprise settings today.

AI at this level can take actions on behalf of users. This includes performing tasks, making decisions, and executing plans autonomously. Several startups, such as Devin, the world’s first fully autonomous AI software engineer, are working on Agentic AI. However, OpenAI places this capability after reasoning, indicating that robust reasoning abilities are foundational for effective and reliable agents.

Level 4 AI systems can create new innovations. This involves not just solving problems but also coming up with original ideas and inventions. Such AI could revolutionize fields like science, technology, and engineering by pushing the boundaries of human knowledge and creativity.

The goal of AGI is Level 5, where AI can perform the work of entire organizations. This level of AI would be capable of managing complex processes, making high-level decisions, and coordinating large-scale operations. Achieving this level would mark the realization of AGI, where AI surpasses human capabilities in the most economically valuable tasks.

According to a recent Bloomberg report, OpenAI is close to reaching Level 2. This development is crucial as it will enhance the accuracy and reliability of AI systems, making them more useful in various applications. OpenAI’s focus on reasoning before advancing to agentic capabilities suggests a strategic approach, emphasizing the importance of solid foundational abilities.

OpenAI’s five-level framework provides a clear roadmap for tracking AI progress. It offers a structured way to measure advancements and understand the implications of each step towards AGI. As AI continues to evolve, these levels will help guide the development of increasingly sophisticated systems.

OpenAI’s commitment to achieving AGI is not just about technological advancements; it also involves ensuring AI’s safe and ethical development. The company’s charter states that if another project is close to building AGI, OpenAI will collaborate rather than compete, emphasizing the importance of value alignment and safety.

While the timeline for achieving AGI remains uncertain, with estimates ranging from a few years to several decades, OpenAI’s structured approach provides a sense of direction and purpose. The journey towards AGI will undoubtedly involve overcoming significant technical and ethical challenges, but the potential benefits are immense.

OpenAI’s five-level framework for tracking AI progress towards AGI offers a comprehensive and transparent way to understand and anticipate AI’s future. By setting clear milestones and focusing on foundational capabilities, OpenAI is paving the way for developing truly intelligent systems that can enhance and transform various aspects of our lives.

Stay tuned for more updates as we continue to follow the exciting journey towards AGI and explore the profound implications of these advancements for our world.

Quinte, along with its partners, is also working towards bringing level 3 agents for banks, who can help banking customers select and apply for the right CDs and loans. Please reach out to me if you want to have a deeper conversation about this.

Sharad Gupta is a McKinsey alum and GenAI, Data Science, GRC (Risk, Fraud Management, AML, Compliance) product and growth leader in banking. Sharad helps CXOs craft AI and automation strategies by bringing innovative GenAI/ML solutions and products in sales, service, risk, fraud, AML, and operations. Sharad can be reached at sharad.gupta@us.quinteft.com or on LinkedIn at https://www.linkedin.com/in/sharadgupta90/.