In a particularly alarming example of unexpected consequences, the bots soon began to devise their own language – in a sense. After being online for a short time, researchers discovered that their bots had begun to deviate significantly from pre-programmed conversational pathways and were responding to users (and each other) in an increasingly strange way, ultimately creating their own language without any human input.
This is the big one. We worked with one particular large publisher (can’t name names unfortunately, but hundreds of thousands of users) in two phases. We initially released a test phase that was sort of a “catch all”. Anyone could message a broad keyword to their bot and start a campaign. Although we had a huge number of users come in, engagement was relatively average (87% open rate and 27.05% click-through rate average over the course of the test). Drop off here was fairly high, about 3.14% of users had unsubscribed by the end of the test.
I will not go into the details of extracting each feature value here. It can be referred from the documentation of rasa-core link that I provided above. So, assuming we extracted all the required feature values from the sample conversations in the required format, we can then train an AI model like LSTM followed by softmax to predict the next_action. Referring to the above figure, this is what the ‘dialogue management’ component does. Why LSTM is more appropriate? — As mentioned above, we want our model to be context aware and look back into the conversational history to predict the next_action. This is akin to a time-series model (pls see my other LSTM-Time series article) and hence can be best captured in the memory state of the LSTM model. The amount of conversational history we want to look back can be a configurable hyper-parameter to the model.
In 2000 a chatbot built using this approach was in the news for passing the “Turing test”, built by John Denning and colleagues. It was built to emulate the replies of a 13 year old boy from Ukraine (broken English and all). I met with John in 2015 and he made no false pretenses about the internal workings of this automaton. It may have been “brute force” but it proved a point: parts of a conversation can be made to appear “natural” using a sufficiently large definition of patterns. It proved Alan Turing’s assertion, that this question of a machine fooling humans was “meaningless”.
Magic, launched in early 2015, is one of the earliest examples of conversational commerce by launching one of the first all-in-one intelligent virtual assistants as a service. Unique in that the service does not even have an app (you access it purely via SMS), Magic promises to be able to handle virtually any task you send it — almost like a human executive assistant. Based on user and press accounts, Magic seems to be able to successfully carry out a variety of odd tasks from setting up flight reservations to ordering hard-to-find food items.
The idea was to permit Tay to “learn” about the nuances of human conversation by monitoring and interacting with real people online. Unfortunately, it didn’t take long for Tay to figure out that Twitter is a towering garbage-fire of awfulness, which resulted in the Twitter bot claiming that “Hitler did nothing wrong,” using a wide range of colorful expletives, and encouraging casual drug use. While some of Tay’s tweets were “original,” in that Tay composed them itself, many were actually the result of the bot’s “repeat back to me” function, meaning users could literally make the poor bot say whatever disgusting remarks they wanted.
To get started, you can build your bot online using the Azure Bot Service, selecting from the available C# and Node.js templates. As your bot gets more sophisticated, however, you will need to create your bot locally then deploy it to the web. Choose an IDE, such as Visual Studio or Visual Studio Code, and a programming language. SDKs are available for the following languages:
Efforts by servers hosting websites to counteract bots vary. Servers may choose to outline rules on the behaviour of internet bots by implementing a robots.txt file: this file is simply text stating the rules governing a bot's behaviour on that server. Any bot that does not follow these rules when interacting with (or 'spidering') any server should, in theory, be denied access to, or removed from, the affected website. If the only rule implementation by a server is a posted text file with no associated program/software/app, then adhering to those rules is entirely voluntary – in reality there is no way to enforce those rules, or even to ensure that a bot's creator or implementer acknowledges, or even reads, the robots.txt file contents. Some bots are "good" – e.g. search engine spiders – while others can be used to launch malicious and harsh attacks, most notably, in political campaigns.
The term "ChatterBot" was originally coined by Michael Mauldin (creator of the first Verbot, Julia) in 1994 to describe these conversational programs. Today, most chatbots are accessed via virtual assistants such as Google Assistant and Amazon Alexa, via messaging apps such as Facebook Messenger or WeChat, or via individual organizations' apps and websites. Chatbots can be classified into usage categories such as conversational commerce (e-commerce via chat), analytics, communication, customer support, design, developer tools, education, entertainment, finance, food, games, health, HR, marketing, news, personal, productivity, shopping, social, sports, travel and utilities.