Page 10 - Monocle Quarterly Journal Vol 3 Issue 2 Spring

MONOCLE QUARTERLY JOURNAL | DEEP LEARNING
1.1 INTRODUCTION
In Douglas Adam’s The Hitchhiker’s Guide to the Galaxy (1979), Deep Thought is a city-sized super-computer built by a pan-dimensional, hyper-intelligent race of people to determine the answer to “the ultimate question of life, the universe, and everything.” After seven and a half million years, Deep Thought calculates an answer, but warns the curious people who anxiously await the response that they will not like it. The answer, it turns out, is 42. Deeply dismayed by this seemingly arbitrary answer, the pan-dimensional beings urge Deep Thought to make sure that this is indeed correct. The super- computer asserts that it has “checked it very thoroughly, and that quite definitely is the answer. I think the problem, to be quite honest with you,” Deep Thought explains, “is that you’ve never actually known what the question was.”
As a nod to The Hitchhiker’s Guide to the Galaxy, IBM chose to name their first foray into a chess-playing program Deep Thought – the predecessor to Deep Blue, which would go on to beat the world champion Garry Kasparov in 1997. And to some extent, this issue of the Monocle Quarterly Journal also takes its inspiration from Douglas Adam’s comical depiction of an all-knowing machine. What we wish to investigate in this edition more specifically, however, is the true nature of artificial intelligence with the aim of dispelling much of the myth, misnomer and hysteria surrounding this often- misunderstood field of research.
To structure this critical analysis of artificial intelligence – with the aim of better understanding what it is capable of at present, what it is not capable of, and what a future saturated with AI might look like – this issue of the Journal has been separated into three parts. The first of these sections includes a brief
history of AI, charting its development from the 1930s to where it is today, interwoven with a technical description detailing important concepts such as deep learning, neural networks, and backpropagation. This first section will also examine some of the successes AI has had in practical terms, especially with regard to the great potential it holds for the banking industry, which has been charged with the important responsibility of identifying and reporting financial crimes, such as tax evasion, fraud, terrorist funding, and money laundering by sifting through untold amounts of data to predict and pre-empt offences.
In the second section of the Journal, we unpack the complex conception of what intelligence really is. Here we interrogate to what extent artificial intelligence can
Deep Thought calculates
an answer, but warns
the curious people who anxiously await the response
that they will not like it.
really imitate human intelligence by examining how the human brain operates and highlighting certain features of humanity that will be exceptionally difficult for artificial intelligence to replicate, no matter the enthusiasm. One such feature – perhaps the most uniquely human trait of all – is the natural acquisition of language, with all its multifarious complexities and nuances of meaning that we seem to grasp so intuitively. This natural capability of humans, along with the many intricate social interactions that language makes possible, still represents a significant hurdle that artificial intelligence must overcome. Despite
8