Bots are dead, long live AI (Part 5/5)

I, for one, welcomed our new bot overlords

tl;dr: Discussing a framework for authentic intelligent messaging agents

Summing up, authenticity in messaging agents is a manifold concept that includes five characteristics.

Transparent purpose: showcase the intent and decision making in a transparent way to make the agent understandable and predictable. Acting transparently and creating confidence in the interactor through the representation of the creator or originator ensures to keep the conversation under control.

Learn from experience: two-way learning where the agent autonomously learns from cultural, behavioural, personal, conversational or contextual interaction data. The learning process is transparent to ensure that the agent is learning in the right direction.

Anthropomorphize: acting as a persona including values, attitudes, and culture to establish a relationship with the interactor. Building individualised experiences to create trust and something at stake.

Conversational behaviour: incorporating necessary communication strategies which include NLP for dialogue handling and other conversational skills. Adapting non-verbal interaction behaviours or patterns, including intelligent reasoning and decision making helps to maintain awareness of the conversation.

Coherence: Staying up to date with the topics of conversation and relating to previous conversations and experiences between agent and interactor. Being aware of the digital and natural context of the conversation and the conversation partner to build common ground.

However, the five characteristics are interrelated where coherence and learning from experience are the main connectors according to the qualitative analysis. Thereby, keeping track of the conversations connects conversational behaviour and coherence. Further, building common ground aims to generate individualised trust and plays to creating a relationship where something is at stake. Representing a persona is tightly connected to the purpose of the agents as well as the internal sortation of the agent. Thereby, acting predictable and transparent, the conversation behaviour is uncovered. To close the circle, by learning from the conversational behaviour and interactions, the agent can advance its capabilities by learning from errors made while interacting as well as by learning from successful conversations. Cultural and conversational awareness both create data to advance the behaviour of the agent and further to develop its persona as well as the confidence of the interactor into the authentic messaging agent

This post is the fifth in a series of five extracts from my Master’s thesis revolving around socially intelligent artificial intelligence. Thanks for reading! :) If you enjoyed this article, hit that ❤ button below. Would mean a lot to me and it helps other people see the story.