From Lab to Life: Does AI Lack a Soul?

Source:  depositphotos.com

On September 8,  OpenAI reported that its AI algorithms had together solved a long-unsolved mathematical problem – one that had eluded the best human mathematical minds.

The problem was called the Navier-Stokes equations problem. These equations describe the flow of air or water or any liquid.  The unsolved question was:  Do the equations permit a ‘singularity’ – that is, a ‘blow up’ or chaotic event, i.e. the liquid suddenly evaporates, expanding and causing damage – or does the system always converge to stability, according to Navier-Stokes?

AI’s answer:  Yes. The equations do permit a singularity. There CAN be chaos.  Proof of concept involved solving an equation where the solution involved a singularity.       OpenAI has since solved some 350  ‘open’ math dilemmas, astounding the mathematics world.

Speaking on the Science Friday podcast,  eminent University of California mathematician Terrence Tao weighed in on the world of math in the AI age.  He made several key points.

We – in fact, nobody, even those who constructed the OpenAI algorithms – know how they reached the solution, he notes.  On many occasions, AI experts have asked AI itself to explain HOW it reached its answers.  It is unclear whether this was done regarding Navier-Stokes, probably not.  Tao claims math is not about finding answers but the way mathematicians define questions.  AI is not expert at that.  Far from it.

In his recent popular book Six Math Essentials, Tao explains why math is so vital for human progress.  Math’s six central concepts show: a) how numbers are the gateway to quantitative thinking; b) how algebra is the gateway to abstraction; c) how geometry is a way to calculate beyond what we can see; d) how probability is a tool to navigate uncertainty with rigorous thinking; e) how analysis is a means to tame the very large or the very small; and finally, f) how dynamics is the mathematics of change.   So, if we want to quantify, project, imagine, assess risk, analyze systems and quantify change, we need math!   AI can help, but always as collaborator.

In his Science Friday interview, Tao laments the decline of humanistic studies among scientists and mathematicians.  He personally regrets not studying more literature and philosophy.  This is a key insight regarding AI.  AI is much like a human with an enormous brain – and without the tools to know how to use it, in order to (in the words of the Bible, Genesis) “to be a blessing for mankind”.  AI’s designers have not figured out how to give their algorithms a soul.  Nor perhaps do they want to, because that might limit AI’s profit margin.

AI is a brain without a soul.  We have had humans like that, and they proved capable of unimaginable destruction – and still are.