The year 2023 may be seen as a significant moment in the rise of artificial intelligence (AI), but its ultimate impact remains uncertain. In February, ChatGPT became the fastest app to reach 100 million users, followed by similar AI models from major tech companies like Google, Amazon, and Meta. These models have the potential to revolutionize fields such as education and healthcare. However, the potential dangers of AI were highlighted in May when leading researchers emphasized the need to prioritize mitigating the risk of AI causing extinction, alongside other global risks like pandemics and nuclear war.
In response to growing concerns about AI, 27 nations, including the UK, US, India, China, and the European Union, pledged cooperation at an AI Safety Summit in November. The goal is to ensure the safe development of AI for the benefit of all. Researchers are focusing on AI alignment, which involves making sure AI models are consistent with human values and goals. However, a major challenge is that large-scale AI models are so complex that they are like black boxes, making them difficult to fully understand.
The lack of transparency and explainability in AI systems is a significant research goal, but it may struggle to keep up with the rapid pace of innovation. This contributes to the wide range of beliefs about AI, ranging from utopian to apocalyptic scenarios. It is crucial that AI alignment works both ways: not only should AI models align with human intentions, but our beliefs about AI should also be accurate. Our beliefs can influence how we interact with AI and how AI adapts to us, potentially increasing the likelihood of different outcomes.
To understand the basis of AI, we need to delve deeper into its origins. AI is based on computation, which traces back to efforts to define arithmetic in logic. However, arithmetic itself is rooted in the deep structure of perception. Recent research has shown that when the limitations of perception are taken into account, arithmetic naturally emerges. This suggests that mathematics is a reflection of the mind and its creativity. While the deep structure of perception is shared with other animals, only humans have invented mathematics, making it a unique and significant creation.
This understanding of arithmetic’s origin aligns with the views of philosopher Immanuel Kant, who argued that our knowledge of the world is structured by innate intuitions of space and time. Kant’s philosophy, known as a “Copernican revolution of the mind,” highlights how the mind shapes perception. This perspective has influenced contemporary psychology and can lead to a similar shift in our understanding of computation. Computation is not separate from us but inherent in the nature of our minds. The mind is more than just computation, and the brain is not a computer. Computation, like mathematics, is a symbolic expression of the mind’s nature and creativity.
What does this mean for AI? Firstly, AI will never become sentient and transcend its artificial nature. The idea of uploading our minds to the cloud for immortality is purely fantasy. However, if the principles of mind on which AI is based are shared by all humanity, we may be able to overcome our individual limitations. Universal computation allows us to simulate and create any outcome in our increasingly connected virtual and physical worlds. AI can be seen as a Promethean fire, a gift to humanity that can be both beneficial and destructive. While efforts to ensure AI alignment, safety, and security are crucial, they may not be enough without awareness and collective wisdom. We need to wake up from the dream and recognize the reality and power of our minds, just like Alice in Wonderland.
