# Toward Next-Generation Artificial Intelligence: Catalyzing the NeuroAI Revolution

**Authors:** Anthony Zador, Sean Escola, Blake Richards, Bence Ölveczky, Yoshua Bengio, Kwabena Boahen, Matthew Botvinick, Dmitri Chklovskii, Anne Churchland, Claudia Clopath, James DiCarlo, Surya Ganguli, Jeff Hawkins, Konrad Koerding, Alexei Koulakov, Yann LeCun, Timothy Lillicrap, Adam Marblestone, Bruno Olshausen, Alexandre Pouget  
**arXiv:** [2210.08340](https://arxiv.org/abs/2210.08340)  
**Date:** 2022-10-15  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2210.08340v3/

## Key points

- Neuroscience has played a crucial role in driving progress in AI for many years
- There is a need to invest in fundamental research in NeuroAI, including the embodied Turing test
- The embodied Turing test challenges AI animal models to interact with the sensorimotor world at skill levels similar to their living counterparts
- This shift focuses on capabilities shared by all animals and inherited from over 500 million years of evolution rather than uniquely human or well-developed abilities like game playing and language
- Most engineers and computational scientists are unaware of the historical importance of NeuroAI or argue that it has lost relevance
- To achieve animal-level common-sense sensorimotor intelligence, researchers must learn from animals and their solutions evolved to behave in an unpredictable world
- Significant resources must be deployed in three main areas: training a new generation of AI researchers equally proficient in engineering/computational science and neuroscience; creating a shared platform capable of developing and testing virtual agents that can pass the embodied Turing test; fostering standardization through common APIs and support for competitions as has been an important impetus for much progress in machine learning and robotics.
- Virtual organisms that demonstrate successful recapitulation of behaviors can be adapted to solve real-world problems after additional efforts are made towards robotics deployment.
- The success of a NeuroAI research program depends on the formation of a community of researchers for whom the raison d'être of their training is to exploit synergies between neuroscience and AI.
- Explicit design of new training programs can ensure that the NeuroAI research community reflects the demographics of society as a whole and is equipped with ethical tools needed to ensure that AI development benefits society.

## Summary

The field of neuroscience has played a crucial role in driving progress in artificial intelligence (AI) for many years. To accelerate progress further, there is a need to invest in fundamental research in NeuroAI which includes the embodied Turing test - a core component that challenges AI animal models to interact with the sensorimotor world at skill levels similar to their living counterparts. This shift focuses on capabilities shared by all animals and inherited from over 500 million years of evolution rather than uniquely human or well-developed abilities like game playing and language. Despite its historical importance in shaping the field of AI, most engineers and computational scientists are unaware of it or argue that it has lost relevance. However, if AI aims to achieve animal-level common-sense sensorimotor intelligence, researchers must learn from animals and their solutions evolved to behave in an unpredictable world. To achieve these goals, significant resources must be deployed in three main areas: training a new generation of AI researchers equally proficient in engineering/computational science and neuroscience; creating a shared platform capable of developing and testing virtual agents that can pass the embodied Turing test; and fostering standardization through common APIs and support for competitions as has been an important impetus for much progress in machine learning and robotics. Ultimately, virtual organisms that demonstrate successful recapitulation of behaviors can be adapted to solve real-world problems after additional efforts are made towards robotics deployment. The success of a NeuroAI research program depends on the formation of a community of researchers for whom the raison d'être of their training is to exploit synergies between neuroscience and AI. Explicit design of new training programs can ensure that the NeuroAI research community reflects the demographics of society as a whole and is equipped with ethical tools needed to ensure that AI development benefits society.