We live in a century of social media and internet. Here, VR (virtual reality) AI (artificial intelligence), fake personalities are almost as common as real people. But how do identify who we are communicating with? What are the habits, patterns, performative elements that we can establish in order to proof that a person on another side of the computer is actually a person?
Working in digital agency, while developing web sites, we often implemented messengers, which are a source of communication between the client and the company. Usually, messengers are set with basic questions or are coded in a way that helps them find words that they are familiar with and propose a reply that fits customer’s question. If messenger’s bot came across a question that he was unfamiliar with, he would propose to call a hot line and clarify inquiry with a live assistant. Messengers created by our agency were usually named after someone (just as Siri) and contained a part of artificial intelligence. By that, I mean they were learning something new from the customer and improved themselves automatically. For instance, if one question was new to them, they would analyze it, address additional questions, propose novel solutions/pages of the website/links and then get to point where client was satisfied. By this, AI (or Alisa, or Ann, or any other Bot-messenger), would update their system and be ready for same question in the future. Agency was not involved in future updates of the system, since it was improving itself. In this particular case, I assume customers knew that they were dealing with a system, rather than with a human, as additional question were asked. Some may not even make sense for the customer, but the system was following its own algorithm, which ultimately solved the case.
Another case that is very common in Kazakhstan – are bots that are following political blogs, news portals and other informational media. Below any news or posts anyone can leave a comment in regard. Those bots are pro-president and are commonly known for typing messages very similar manner. Also, they appear only in those materials which (by government opinion) need correction, opposition, support of parliament.
If we talk about how to determine whether I am talking through what’s app, facebook messenger, Skype, other, with a real friend of mine or his computer duplicate, perhaps in the future there could be some kind of a finger print detector, which initiates the conversation?
Did you, in your encounters with people, ever have to define yourself as a person rather than a bot? I mean, did your company mark when the system switched from bot-messenger to real person? Should they?
ReplyDelete