Waiting for a response for an online enquiry, trying to cancel a subscription for a magazine i never read; I was perhaps already slightly frustrated. I found myself infuriated about not knowing if I need to respond to the pleasantries coming up in the text window from “Katie” or if I can just ignore them? Is this just an overly friendly clerk at a desk, or a piece of code programmed to seem as human as possible. Then I wondered “why do I care so much about this? Does it really matter?”
Humans have always seemed to have a fascination with robots. Playing god does sound pretty appealing. One of the early science fiction stories, from stone golems and living puppets, is a being in our own image. Creating something out of nothing, is like having a child, without all the mess, and with the potential to outlive all of us.
One of the cornerstones of the evolution of the human species was our ability to develop tools; from these simple tools, more complicated machines emerged, but complexity does not equal intelligence. Early machines were designed to do one job, and one job well. Cogs, steam and springs were the pinnacle of technology for a long time, but we were still dreaming of mechanical men. Perhaps the most famous example of this was The Mechanical Turk, a chess playing automaton, built by Wolfgang von Kempelen in the 18th century to impress local nobles. And it did impress, beating many famous thinkers of the time, from Napoleon to Benjamin Franklin. However, this may be the start of a common thread throughout this article of distrusting machines. The Turk turned out to be an elaborate illusion, with a chess master hiding within its working, pulling all the strings. Although now chess playing machines have come on leaps and bounds from the Turk, with systems that outplay even the best of human players. Some would argue that this “A chess playing machine that plays chess well” still does not show true intelligence!
When it come to artificial intelligence, most people think of one person, Alan Turing, and although his contribution to this field cannot be understated, another person who deserves a mention is Joseph Weizenbaum, a computer scientist working for MIT, who created a program in 1966 which he called ELIZA. ELIZA was capable of engaging in conversations with humans, asking simple but reassuringly open questions which would resemble what a psychologist might ask, and it was hugely effective, with ELIZA’s first test subject, Weizenbaum’s secretary, asking him “would you mind leaving the room please?” so that she could continue her conversation with this primitive chatbot.
There are obviously big gaps still present within the computer’s vocabulary; humans have accumulated vast amounts of subconscious shared knowledge which we use to plug these holes between thoughts and phrases, which a computer finds hard to replicate, creating an uncanny and clumsy perception of our world. However, another aspect of the human mind which can be forgiving of these social foibles, is our ability to empathise with others, even if those others only slightly resemble a human being (possibly to their advantage-see the uncanny valley). The simple act of giving a robot a name or placing googly eyes in an otherwise impression less face can drastically alter the way humans treat the robots and computers we share the world with. We can start attributing the actions of a robotic system to conscious thought and consideration; this anthropomorphisation is also known as the ELIZA effect.
Without the work of Turing and Weizenbaum it might be argued that we would not be where we are right now with computing and AI. However, they both shared a worry that these computer systems could be used to fool people and would be exploited nefariously by some. In fact, throughout the later part of Weizenbaum’s career, he would publicly speak against the development of these systems, fighting against the machines he had an integral part in creating. With the onset of the ubiquity of the internet, these chatbot systems have become much more advanced, having infinitely more examples of speech to hone their mastery of language and with GOOGLE, the font of all knowledge to refer to. And while yes, all tools can be used for good or for ill, there has been a massive uptake in the use of these chatbot systems for a societal good. Going all the way back to ELIZA, chatbots have been developed to give people access to support in areas of mental health deserts, as well as developing innovative solutions for distributing important information.
While still waiting for a response from “Katie” (a computer wouldn’t take this long, would it? Or is this some elaborate double bluff? Ow my brain….). Is this what philosopher, John Searle, talked about when he discussed “The Chinese Room Argument”? does this ambivalence equate to passing the Turing test? is this what true Artificial Intelligence looks like? Faking it just enough to scrape by…. That does sound very human. Weizenbaum once used the analogy of a dinner party when describing modern chatbots. You are introduced to someone new, who describes what their job is, you have no idea what this means, but using what they have told you and social cues you have collected in the back of your brain, you are able to maintain the conversation just enough to not make too much of a fool of yourself, just about. “It’s a con job” as Weizenbaum describes it, a cold read, as much of what chatbots do is just mirroring what you say, finding keywords and linking this back through its archives to try and find an answer to what you are referring to. Although, as I type this, that just sounds like most conversations. This asks the question, what is intelligence? Is the human mind just a machine that is good at making connections and linking them to a database of reference, something it does really well? And perhaps we get what we deserve, we trust computers enough to let them in all aspects of our lives, perhaps they deserve a little bit of common courtesy from us.