Botch, bother, botulism, botox. Have you noticed how many unpleasant words start with a bot? Unfunnily enough, that’s more or less the story of my week: a string of unpleasant words and unhappy experiences starting, and, it seemed for much of the week, ending with a bot. This week’s post is about one particular aspect of AI, which, I remind you, ostensibly stands for artificial intelligence. However, I have my doubts. I have spent far too much time this week in conversation with a bot whose sensitivity to language – intonation, nuance, register – were so sophisticated as to leave me in doubt whether it was artificial, but, sadly, the limitations of whose algorithms left me in no doubt that it was not intelligent.
I’m writing this midday on Monday, which feels like about two weeks since the week began, and I currently don’t know how this story ends. The upside of that is that I already have a topic for next week’s post: David’s Adventures in Botland: Part II. The downside is that, as of this moment, I don’t know how traumatic Wednesday will be.
It will certainly be traumatic: in my now considerable experience, an MRI is not a pleasant way to spend a Wednesday afternoon. The double prospect of the claustrophobia of being glided into a narrow tube, and the ridiculous cacophony of loud noise that bombards you have me tensing up already. However, as will become clear as we go on, if that is what constitutes my trauma this Wednesday, I will regard it as a good day.
One of my specialists has referred me for an MRI. The technical procedure is that when my health fund has issued a commitment to pay for the test I can make an appointment for it. Knowing that there was likely to be a long waiting list. I anxiously awaited news of the issuing of the authorization. As soon as our local clinic phoned me with the good news, last Thursday, I phoned the appointments number for the institute that carries out the test. At the other end of the phone was a bot: quite a sophisticated bot with an impressive armoury of increasingly exasperated tones of voice when it does not get its way. I carefully explained why I was phoning and the bot informed me that the Institute does not perform such a test. Since I knew this to be inaccurate, I held my ground.
As the bot was having nothing of it, I asked to speak to a human being. The bot then pointed out that it could only put me on to a human being when it was clear that it was unable to handle my request. From the bot’s point of view, since the institute didn’t perform such a test, the bot had handled my request. It therefore refused to pass me on to a human being. There is only so long you can go on arguing with a bot, although in my experience it’s a little longer than you should bother with. In the end, it was the bot who said, in a very reasonable simulation of a huff: “I’m going to have to conclude this conversation now. This will take a moment.” I allowed myself the satisfaction of hanging up before the bot’s moment was through. I’m damned if I’m going to let a bot hang up on me.
Obviously what I needed to do was to find a way to get through directly to a human being at the institute. A thorough online search of their site and various other avenues revealed nothing more than the one phone number which only leads to a bot. I therefore phoned my own clinic where after initially suggesting that I phoned the institute, the clock soon agreed to send a message to the institute asking them to contact me.
On Sunday, as I was still trying to figure out a way to actually get through to a human being, i received a phone call from the institute to tell me that they had received the authorization from my health clinic and the very pleasant clerk at the end of the line told me that he was phoning to make an appointment I took the liberty of explaining the problem with their bot to him before making an appointment, which was probably a tactical error, although he seemed very appreciative of the fact that I was making him cognizant of the limitations in the bot’s algorithm.
Within a matter of minutes, we had made an appointment which was, to my absolute astonishment, not for three months’ time as I had suspected would be the case, or even three weeks’ time, as I had hoped, but rather three days’ time. Moreover, it was for an eminently reasonable time of day 2:00 p.m. I thanked the clerk profusely and waited to receive the SMS confirmation of the appointment with all of the details of how I had to prepare for the test
Within minutes the confirmation arrived. However, the SMS referred to an MRI subtly but very significantly different from the one that I was authorized to have and that I had made an appointment for with the clerk. Indeed, the test specified was about half of my body distance away from where it should have been, rendering it more or less absolutely superfluous.
I immediately called the number from which the clerk had called me; none of you will be surprised to hear that a recorded message informed me that this number was out of service. Pausing only to pour myself a large whisky, I phoned the appointment telephone number of the institute and spoke to the bot. I carefully explained what I needed and the bot explained with equal care just what my options were: if I was calling about an existing appointment, I could change the date, the time or the location, but not the type of test. In a classic demonstration of the triumph of optimism over bitter experience, I asked the bot to put me through to a human being. I leave you to complete the rest of the conversation.
In addition, having finally found online an email address for the institute, I sent an email explaining the situation and enclosing both the authorization from my health fund stating clearly the nature of the test and a screenshot of the SMS stating clearly the nature of a completely different test for which I had received confirmation from the institute. I wonder if any of you have received a phone call from the institute or a response to my email, because, having just checked, I can tell you that I have not. In fact, after a brief discussion, Bernice and I have decided that driving to the institute tomorrow, Tuesday, to sort this out in person on the day before the test, is a marginally less depressing prospect than turning up on Wednesday for a test which I may be told I cannot take because they have prepared for a completely different test. Rather than my typing this week’s post on Tuesday I am dictating it into my phone as Bernice and I drive to Zichron Yaakov on Monday in the hope of leaving myself free enough on Tuesday to be able to waste three or four hours going to a far from central part of Jerusalem to try and get this sorted out so that I can actually have the test on Wednesday, weeks, if not months, before I dreamed I would be able to have it.
I promise next week to bring you up to date, however depressing or delightful the update proves to be. [Spoiler alert: as of Tuesday evening, things look very promising.]
Meanwhile, I really hope we all have a good week