Chatbots - the next big thing?

Turing Test

The history of Chabot’s begins with the inception of early variations of natural language processing. The history of natural language processing as a field of Artificial Intelligence development began throughout the 1950s. In 1950, the famous Essay by Alan Turing "Computing Machinery and Intelligence" (what is now dubbed as the Turing Test) first proposed a criterion for judging whether a computer program that can communicate with humans—a Chatbot—is intelligent enough to sufficiently fool a real human judge into thinking the Chatbot is a human actor. In 2014, the Chatbot Eugene Goostman from Russia was the first of its kind to pass the Turing Test; successfully convincing 1 out of 3 judges it was 13 year-old non-native English speaking, Ukrainian boy.

The Current Situation

Most future predictions in technology point towards Chatbots advancing from their current form, into more intelligent, multi-faceted and being able to register complex aspects of language and human cognition than they currently possess. However, there is not a complete general consensus on the viability of Chatbots for various uses. Some commentators point out some shortcomings of current Chatbots, and their inability to tackle complex situations — like those involving emotions or irony for example.

Recently, Microsoft suffered a failed experiment with its social media based AI Chatbot ‘Tay’ — who was designed to learn and replicate the personality of a 19 year old girl, by interacting with people on social media platforms Twitter, Kik and GroupMe. Tay started off as harmless and innocent as to be expected — talking of love and puppies — but soon, her behavior descended into something of a racist, holocaust denying, fascist. The cause was manifold—for one, a number of shortcomings existed in Tay’s code that meant for one — she lacked the ability to reflect on the remarks and notions presented to her. Unlike normal 19 year old women, Tay could not make a personal judgment to distinguish between truth and untruth, propaganda and opinion from fact. To this end she was exploited by trolls that took advantage of a feature that would cause her to basically repeat phrases said to her when requested.

While others theorize Chatbots taking a dominant and role in the modernizing of technical and business customer service functions across all aspects of technology: with Chatbots engaging in problem solving and customer service conversation, equipped with the ability to complete multiple transactions, or detect the appropriate language of the user for example — automating tasks that otherwise do not necessitate a human actor.

Future Directions

The development of Chabot’s sits within the grasp of the broad advancement of Artificial Intelligence (AI) systems — that is a multidimensional field of research and development that extends through the meeting point of many disciplines. Current predictions for the development of Chabot’s in the coming years are intertwined with the notion of the Post-App Era—a term coined in the top 10 predictions of the 2016 Gartner Symposium. The Post-App era can be categorized as not the end of app use in itself, but an evolution in how humans interact with applications, and the diminishment of the single-use app; apps that users download or have for a single purpose, for say banking, dating or chatting.

By 2020, smart agents will facilitate 40 percent of mobile interactions, and the postapp era will begin to dominate - Gartner

Today most of us tend to only use several apps on a daily basis, most common of which came preloaded with their phone. The Post-App era envisions a time when the user does not need to open separate individual apps, but rather, a single functional app that completes the requests of users — similar to Apple’s, Siri—but with more intelligence, learning ability and agency than she is equipped with today. Technologists term these apps, Smart Agents. Other examples of existing smart agents include computerized personal shoppers. Chatbots are, and will be the voice of smart agents, for example—things automatic email clients fit this category.

Smart Agents

Decades of work have gone into the development of the modern conversational agent or Chatbot, and this technology is an integral component of advancing artificial intelligence systems. Chatbots are a spawn of the development of natural language processing systems in AI—and the movement towards robots having complex speech recognition. Facebook employing the use of Chatbot technology to deal with customer queries is a current example of this. A smart agent is a computerized agent that is able to fulfill cognitive tasks, adapt, communicate and learn. s computers are embedded in more and more everyday objects around us, the objects themselves become smarter. As such the space around us is set to become more and more interconnected as the potential for the omnipresent interconnectivity of all these intelligent objects — the ecology of smart agents — becomes an ever growing reality.

The impact of smart agents is predicted to impact the organization, as much as the personal. Technologists are only beginning now to theorize on the impact having a growing number of artificial smart agents performing functions in an organizational framework. One theory is that smart agents will speed the diffusion of information. Smart agents are capable of of providing work within, and among different organizations.