AI Expert Rodney Brooks has recently been vocal about his doubts on OpenAI’s large language models, claiming they are far from being able to reason at a level to compete with humans. OpenAI’s most popular application, ChatGPT, is based on these language models. Through an enlightening interview with IEEE Spectrum, Brooks questions whether AI should be seen as a worthy opponent to human intelligence, which he dubs Artificial General Intelligence (AGI).
Brooks’ argument is that these tools are being greatly overestimated for their capability. He believes that these models do not contain any comprehension of the world outside of linking correlations found in language. This implies that when a person interacts with an AI tool, it is a result of searching for those language correlations as opposed to rational thought.
Brooks’ opinion on AI also intersects on his attempts to help himself with coding. While he found ChatGPT to be useful, he noted the discrepancy between the tool’s confidence in each response and how wrong it turned out to be at times. He states that a model’s performance does not equate to its level of intellect.
Overall, Brooks believes advanced AI tools can still be beneficial, but they won’t be able to make the jump to AGI anytime soon. His stance is an important reminder that we should consider, especially in an age where the future of AI is increasingly uncertain.
OpenAI is a cutting-edge research lab run by a team of world-renowned scientists and industry figures. Founded in 2015, OpenAI is a non-profit research lab dedicated to tackling the long list of challenges associated with developing AI—primarily the development of AI systems that we can trust, understand and control.
Rodney Brooks is an emeritus professor of Robotics at the Massachusetts Institute of Technology (MIT). He is also the founder and chairman of Rethink Robotics. He has won numerous honors in robotic research and development, in addition to being elected to the National Academy of Engineering. Brooks is well respected in the AI research community, and his conversation with IEEE Spectrum highlights the importance of keeping a healthy level of skepticism towards AI advancements.