Foreword
"How Far is Too Far?" reads the title of the first episode of the newly released Youtube Originals Documentary "The Age of A.I." where nobody other than Robert Downey Jr., known through his outstanding performance as Tony Stark and Iron Man in the Marvel movies, narrates current developments, trends and state of the art applications in the field of A.I. And while Robert Downey Jr. tries to explain Artificial Intelligence, Machine Learning and Artificial Neural Networks one cannot get away from the fact that the rapidly evolving field of data science has not only hit mainstream media but started to gain an increasing amount of interest from the general public. However if one thing is for certain, then that this polarizing topic has not only been discussed too much but most importantly it has been discussed the wrong way!
Misunderstood Intelligence
"Artificial Intelligence" is a very misleading term, not only because there are several different definitions for it, but especially because nothing in current Artificial Intelligence actually is intelligent. Technically Artificial Intelligence is just an umbrella term for a collection of mathematical models that are capable of finding stochastically sound decisions or values based on underlying data. Most prominently among those models is probably the so called Artificial Neural Network, often abbreviated as just Neural Net. Another misleading term and often misunderstood by data scientists without biological background, the Artificial Neural Network is not a model of the brain and neither is it the model of a human neural network, even though it has been conceptually copied from it. As a matter of fact an Artificial Neural Network is nothing different than a chain of mathematical functions that are optimized on some input data. However putting it that way is a lot less spectacular than saying that our computers started to think and possess some kind of intelligence (or maybe even a brain). Today it sadly is common practice to exaggerate and claim the latter even in scientific research, as buzzwords have become increasingly more popular and important for funding of research projects. Whether or not this approach is moral and valid or if it defeats the purpose of scientific research is debateable. However it gets extremely frustrating when study material, documentations and even universities start to teach these matters wrongly with such conviction.
Mathematical Alchemy
"Machine Learning has become alchemy" notes Google researcher Ali Rahimi in his Test-of-Time award speech at NIPS 2017 as he talks about how Artificial Intelligence has yielded impressive results, from self-driving cars being around the corner, to automated tagging of faces in pictures over to transcribing and translating voicemails and documents. For these kind of tasks alchemy is fine, he adds, "But we're now building systems that govern health care and our participation in civil debate. We influence elections. I would like to live in a world whose systems are build on rigorous, reliable, verifiable knowledge, and not on alchemy." [1], [2].
The Edges of Explainability
It may come as a surprise - especially since I try to explain everything, all the time, with mathematics - but I do actually think that even mathematics and its applications has its limits. Maybe it is due to the saying that mathematics is the language of nature that we tend to believe that we are able to model everything with it. But contrary is the case, nature does not speak maths, we invented mathematics to describe the complex processes that happen in our universe. Similar to the fact that nature does not abide by natural laws but the other way around, our natural laws being a collection of patterns that we observed and learned to understand over time. I think that some things are too complex and too fantastic to model them mathematically: Feelings, conscience, intelligence... Concepts that we ourselves do not even fully understand but claim to be able to recreate.
But maybe I am just a prisoner of my own world view, to quote Andrew Ryan from the game Bioshock: "Have I become so convinced by my own beliefs, that I have stopped seeing the truth?" Maybe I just cannot accept that something as wonderful and incredible as the human brain, even humans itself, can be broken down to simple mathematical functions. Maybe it is just a wish, the desire that there is more to us...