Recently, I had the opportunity to join a discussion held by a group of internationally renowned scholars working on the brain, focusing on applications of neuroscience and artificial intelligence. Since I had already started writing on this topic, I’d like to briefly share some of the notes I took during the conversation.
First, it’s worth noting that although expectations around artificial intelligence are quite high, there are still significant differences of opinion even among leading scholars striving to understand the brain and its functions—particularly on whether true AI can possess human-like intelligence, and if so, how that might be achieved. While exploring the question of how AI should be, I encountered various perspectives and questions that were genuinely exciting. Below, I share a few highlights from my notes.
For instance:
Although intelligence is often associated solely with the brain, one notable viewpoint argued that the body also possesses its own form of intelligence (and no, they weren’t talking about the gut this time! 😊).
The importance of understanding the sensorimotor system in developing human-like AI was emphasized. The summary point was this: to create an intelligent agent, one must understand the body—its muscles, joints, and embodied intelligence. Because adaptation and learning inherently involve grasping both the brain and the body.
Some researchers even suggested that, in certain respects, our motor system may be “smarter” than the brain itself. This idea draws from embodied cognition theories, which propose that the body is not just a passive carrier controlled by the brain, but an active participant in cognitive processes. According to this view, muscles, joints, and sensory organs don’t merely execute brain commands—they contribute directly to the formation of intelligent behavior through interaction with the environment. Developing true AI, then, may require a deeper understanding of this system and its dynamic relationship with the brain.
The neuroscience community has vast amounts of data on the brain, but this data doesn’t necessarily explain how things work. Neuroscience is largely a theoretical field, while AI serves as a testing ground for many of its hypotheses.
Some of the thought-provoking questions that caught my attention included:
- How can we define the traits that make humans intelligent? What are these competencies? What kinds of tasks highlight these capabilities and allow us to measure them?
- To what extent can the field of neuroscience provide concrete insights into building better AI?
One of the most important problems raised was the challenge of continuous learning: How can we train an artificial agent to continuously improve over time?
A question I’ve been curious about also came up: Does intergenerational learning transfer exist? There were varying views. Some scholars argued that development begins only after birth, while others suggested that certain genes—especially those involved in brain development—might transmit the effects of previous generations’ experiences. Rather than proposing a direct Lamarckian inheritance of knowledge, this view is grounded in epigenetic: the idea that environmental influences can modify gene expression through chemical mechanisms, and that some of these modifications may be passed on. This sparked further debate on whether epigenetic inheritance could influence cognitive predispositions across generations.
The questions and challenges are many, but I hadn’t felt this level of academic excitement in a long time. It was a refreshing and motivating experience.
Note: This is a translation of a Turkish post from 2022
Dr. Olgun Sadik's Personal Website 
Leave a comment