# If consciousness is dynamically relevant, artificial intelligence isn't conscious

**Authors:** Johannes Kleiner, Tim Ludwig  
**arXiv:** [2304.05077](https://arxiv.org/abs/2304.05077)  
**Date:** 2023-04-11  
**Summary page:** https://www.summarizepaper.com/en/paper/2304.05077v1/

## Key points

- Consciousness and artificial intelligence (AI) systems are explored in the paper titled "If consciousness is dynamically relevant, artificial intelligence isn't conscious" by Johannes Kleiner and Tim Ludwig.
- The authors argue that if consciousness affects a system's temporal evolution, AI systems cannot be considered conscious.
- AI systems operate on processors like CPUs, GPUs, and TPUs that adhere to computational dynamics designed to prevent deviations.
- These design and verification processes eliminate potential consciousness-related dynamical effects in AI systems.
- The authors suggest that AI systems lack dynamic relevance, an important aspect of determining consciousness.
- Consequently, the study concludes that AI systems cannot possess consciousness.
- The research highlights the limitations of current AI technologies in replicating human-like consciousness.
- Existing AI technologies fall short compared to humans in terms of conscious awareness.
- The findings contribute to ongoing discussions about the nature of consciousness and its relationship with artificial intelligence.

## Summary

In their paper titled "If consciousness is dynamically relevant, artificial intelligence isn't conscious," authors Johannes Kleiner and Tim Ludwig explore the relationship between consciousness and artificial intelligence (AI) systems. They argue that if consciousness plays a role in the temporal evolution of a system's states, AI systems cannot be considered conscious. The authors highlight that AI systems operate on processors such as CPUs, GPUs, and TPUs which are specifically designed and verified to adhere to computational dynamics that systematically prevent or suppress deviations. These design and verification processes effectively eliminate any potential consciousness-related dynamical effects. By emphasizing the importance of dynamic relevance in determining consciousness, the authors suggest that AI systems lack this crucial aspect due to their predetermined computational dynamics. Consequently, they conclude that AI systems cannot possess consciousness. This study sheds light on the limitations of current AI technologies when it comes to replicating human-like consciousness. It provides insight into how existing AI technologies fall short in comparison to humans when it comes to possessing conscious awareness. The findings contribute to ongoing discussions surrounding the nature of consciousness and its relationship with artificial intelligence by demonstrating how current AI technology fails to meet certain criteria for being considered conscious.