Berkeley Alum John Schulman has had an impressive journey with Artificial Intelligence. His major career achievements include cofounding the software company OpenAI and leading the reinforcement learning team that developed ChatGPT – a chatbot based on Generative Pre-trained (GPT) language models. A global sensation, ChatGPT has a remarkably human-like ability to generate responses.
Berkeley News spoke with Schulman about his journey and his experiences at UC Berkeley where he graduated with a Ph.D. in Electrical Engineering and Computer Sciences.
Schulman had initially considered pursuing physics, since he admired great physicists like Einstein. However, he was not as interested in the summer research projects in physics, and decided to explore other topics such as Neuroscience. After finishing a lab rotation with Pieter Abbeel, he got excited about Abbeel’s work on helicopter control and towel-folding robots and ultimately switched to the EECS department.
Schulman desired to understand the universe and AI seemed to be the way to do it. OpenAI was attractive to him, since it was aspiring to build Artificial General Intelligence (AGI). He finally chose OpenAI, because it was accepting to talk about AGI. His research focused on Deep Reinforcement Learning, particularly in tying knots with a Personal Robot 2 (PR2).
The paper “Deep reinforcement learning from human preferences” by another Berkeley alum – Paul Christiano – created a pathway for Schulman to explore and which ultimately led to the development of ChatGPT. ChatGPT became very popular among the public because it was easier to use than previously existent models of similar quality and had lower “hallucinations”.
He notes his concern for any potentially harmful use of ChatGPT, yet clarifies that it’s misuse risk is currently not a major concern but that a “treacherous turn” of the AI system is a risk OpenAI is taking seriously. Schulman believes that AI will gradually take over tasks that used to be done by humans, forcing people to rethink their roles and how to use AI to their benefit.
OpenAI is a software company founded in December 2015 by Elon Musk, Sam Altman, Greg Brockman and John Schulman. It is a world leader in artificial intelligence, focused on developing technologies that make the world a better place with human-level AGI. OpenAI’s professional safety team researches ways to align AI with human preferences while avoiding misuse.
John Schulman was born in the United States. He is a co-founder of OpenAI, an ambitious software company and the leader of the reinforcement learning team. Schulman has a Ph.D. in Electrical Engineering and Computer Sciences from UC Berkeley. He worked with Pieter Abbeel and Paul Christiano for projects in robotics, such as tying knots with PR2 robots, and deep reinforcement learning from human preferences. He developed the global sensation ChatGPT, based on OpenAI’s generative pre-trained (GPT) language models. He is passionate about understanding and advancing the field of AI and is conscious of the risks associated with artificial intelligence.
