The Bot Revolution is being talked up as the next big thing.
The word bot derives from robot and first appeared during the chatroom era, which peaked in the pre-Web time around 1985. The word itself goes back to the 1400s but it was a derisive term that had nothing to do with robots. It has since transformed into something miraculous—it's the future.
If you encounter a bot trying to start up a conversation in an IRC chatroom, you can put the word "bot" on a single line and most chatbots will return the name of the bot and who owns it. If you ask "are you a bot?" it will usually return some bogus answer trying to convince you it is not a bot. I consider these legacy bots.
Even more fun are the over-the-phone robo-calls, which are now employing some form of interactive bot that is quite fascinating. I've only encountered three of these in the wild over the past six months. Two came with a female voice similar to the ubiquitous "Rachel from card services" robocaller.
The first time I ran into this robot, I stupidly hung up, passing up a research opportunity. The second time I interacted to observe the quality of the coding. It was decent. The voice was excellent, and it reacted accurately to whatever I said. There was a noticeable lack of real emotion, and there was an obvious fake enthusiasm in the voice. I was never sure what I did wrong, but after a few minutes of interaction it just stopped talking and forwarded the call someone named "Bill," who asked me if I wanted to fix my credit card problems.
When I told him I don't have any credits cards and asked him why did he called, he abruptly hung up on me.
A month later, I finally got my third interactive robot call. This time I was going to see how long I could keep it on the line. It had a male voice and some sort of odd buzzing on the line that sounded like tape hiss. I immediately thought it was a robot.
I should mention that with the other two calls, it took at least three interactions to be sure it was a robot although I immediately got suspicious. I'm guessing the machines are not fast enough to parse what I said and then grab the right clip to play to me. It went something like this:
"Hi, is this John?"
"How are you today?"
"Who is this?"
I discovered that any off-the-wall answer or question made the machine work harder. I'm guessing that after the "how are you?" question is asked, you can have the most fun. I'd suggest the following responses: "Why are you asking me that?" "Let me ask you the same thing." "Do you have a puppy?" You get the idea. You should also experiment with variations on the appropriate answers.
I did manage to get the get the bot into a short conversation. The longest pause came from my blurting out "Are you a bot?" I think two or maybe three seconds went by before it said, "no."
This immediately answered the question, "Do bots lie?" asked by various tech philosophers who actually take these things seriously. From my experience bots always lie. They are programmed to lie.
Lying is a natural condition of bots. So there is no reason to suspect this will ever change, making them unreliable from the outset. This means that all these high expectations for bots to fill in as a cheap personal assistants (ordering Uber rides and straightening out your calendars, getting your plane tickets, etc.) is risky.
Worse, from my perspective, I've been hearing these exact same promises since the 1980s. Only then it was not about "bots." That's when the refrigerator was going to order milk on its own and the washing machine was going to phone a repair man when it needed fixing. The so-called bot revolution is a re-up of these old bullcrap promises. Nothing has changed except perhaps the robo-calls are more fun.