OpenAI has developed a set of five tiers to track its progress in building AI software that can outperform humans, the startup’s latest effort to help people better understand its thinking about security and the future of AI. The ChatGPT creator, widely considered a leader in the race to build more powerful AI systems, shared the new classification system with employees Tuesday at an all-hands meeting, an OpenAI spokesperson said. The tiers, which OpenAI plans to share with investors and others outside the company, range from the type of AI available today that can interact with people in conversational language (Tier 1) to AI that can do an organization’s job (Tier 5).
OpenAI executives told employees that the company currently believes it is at the first level, according to the spokesperson, but is close to reaching the second, which it calls “Reasoners.” That refers to systems that can perform basic problem-solving tasks as well as a human with PhD-level training and no access to tools.
At the same meeting, company executives demonstrated a research project involving its GPT-4 AI model that OpenAI says features new skills that enable human-like reasoning, according to a person familiar with the discussion who asked not to be identified because they were not authorized to speak to the press. Reached for comment on the demonstration, the spokesperson said OpenAI is always testing new capabilities internally, a common practice in the industry. OpenAI has said for years that it is working to build what is often called artificial general intelligence, or AGI — essentially, computers that can do a better job at most tasks than people. Such systems don’t currently exist. OpenAI CEO Sam Altman has previously said he expects AGI to be achieved within this decade. AI researchers have long debated how to determine the criteria for achieving AGI. In a November 2023 paper, several Google DeepMind researchers proposed a framework of five ascending levels of AI, including levels such as “expert” and “superhuman.” These rankings resemble a system often used in the automotive industry to assess the degree of automation of self-driving cars.
According to OpenAI’s tiers, the third tier on the AGI path would be called “Agents,” referring to AI systems that can spend days performing actions on a user’s behalf. Tier 4 describes AI that can come up with new innovations. And the most advanced tier would be called “Organizations.” The tiers were developed by OpenAI executives and other senior leaders, and are considered a work in progress. The company will gather feedback from employees, investors, and its board, and may adjust the tiers over time.
One more thing! We are now on WhatsApp channels! Follow us there to never miss any updates from the world of technology. To follow the HT Tech channel on WhatsApp, click on here to join us now!
