It’s a modest attempt to bring contrasting views of two great minds of the 21st century and explore potential solutions emerging at the intersection of these ideas. Pursuing such integrated solutions has become essential in guiding humanity to adapt to the new realities and challenges of artificial intelligence in the future. On one end of the spectrum, we have Geoffrey Hinton, the Father of Artificial Intelligence, who categorically dismisses the existence of consciousness and has a pragmatic and physicalist approach towards AI and its potential risks for humans. On the opposite end, we have Federico Faggins, an almost Jungian idealist who posits that consciousness is fundamental to everything and our last resort to mitigate the complexities that the development of AI will bring in the future.

Since leaving Google, Hinton has vociferously expressed concerns about the dangers of AI. His major concern stems from his realization that digital computing can be better than analog computation in certain aspects.

BRAIN VS AI MODEL

Our Human brain predominantly runs on analog computation, with hundreds of trillions of densely packed neurons miraculously running on meager energy consumption of only 3 watts. The largest AI multimodal of today has only a trillion connections and requires a substantially larger amount of energy to operate.

At prima facie, this data gives humanity a momentary assurance about AI dangers, suggesting that the human brain will remain superior to future generations of our current AI model. However, it is important to note that Digital AI computation may further encounter limitations and bottlenecks in its development due to energy saturation, chip size limitations, etc.

At this juncture, Hinton comes in with his essential insight that amidst the energy efficiency of the human brain, the AI models have a distinctive feature that humans cannot fathom developing. He calls it “Superior Sharing”. Each AI model running on multiples copies on various hardware in networks all across the globe can share trillions of data faster than any human. Humans share through language, which is a very inefficient way of sharing compared to these AI models.

So Hinton envisages the emergence of a vast neural organism distributed across the globe. The system outperforms humans in capacity and has a superior ability to learn.