It is a rainy day. The dark clouds drift slowly over the city. In a narrow alley the streetlights awake. Dave, a manager in his mid-fifties, looks out of the window outside and wonders: “Streetlights, in the middle of the day?” This weekend is his wedding celebration. He and his wife Susanne have rented a nice little place in the countryside, with a nice veranda. “With this weather…,” Dave turns back to his computer. “Where was I? The numbers from last month compared to the average…” The rain gets louder, and it pounds against the small window. “Argh, the rain, it reaches my neck.”
“Please pay attention to the numbers, the results have to go to your manager within 15 minutes,” says a computerized voice from his desktop. “If the rain is bothering you, I can play some music. Close the window and rebook your cottage.” “What?” Dave still isn’t used to this new computer system, it knows too much.
My Dutch blog of May the 4th (https://laurent.fraters.org/blog/english-blog) is about computer systems being aware of their surroundings. And not only aware of their surroundings, but also being capable to act as and respond as intelligent people would do. Will it be possible in the near future, or is it a fantasy film story? One of my readers provided a link to an article that tells: “Computers are knocking on the door of the company boardroom”. Graham Kendall, who wrote this article, concludes: “Technically it’s possible, but the costs to develop and maintain … probably means that it is not within the reach of most companies … at least for now”. However, Simon Wardley puts forward that Business is nothing more than a catfight. He concludes that if some companies do get an advantage by the use of new technology, others will follow soon or become irrelevant.
Will it work? Will the computer be trusted enough by the other board members to get a real vote? The computer can access and evaluate far more data than any human being ever could do. Comparing all this data and calculating the most relevant answer is one step. To make the computer boardroom vote count, the other (human) board members should really understand how the computer came to its conclusion. In which way can computers convince humans? Partly, humans will trust computers more and more as their lives become dependent on computer systems all along. Additionally, for the more complex and high-risk decisions, as in a boardroom, a computer should be taught to present their results in a way that humans perceive their surroundings.
Humans use metaphors to understand the world around them, as Lakoff and Johnson point out a note in their book ‘Metaphors We Live By’. Metaphors visualize a comparison of known and not known, bringing up some aspects and hiding others. This way, we can understand things or situations we have not seen before. I give you an example that might be familiar if you have little children. Let us assume that the first animal Tessa encounters is a rabbit and she learns that it is called a rabbit. Then, if she sees a cat, she may refer to it as a rabbit with a tail. Over time, she will learn that this new animal is a different type of animal, a cat. Consequently, computers should visualize their results in accordance with the metaphors that would be understood by the other (human) boardroom members.
However, for Dave, this new computer system is like a nightmare; “it will know before me when I get sacked”.