Quality Aspects of Bots

Abstract

Bots are a fascinating species. Although they are generally considered to be a product of artificial intelligence (AT) research, they are sort of artificially block-headed, simply displaying human understanding while not possessing a single glimpse of reasoning. However, the available technical means of software robotics provide amazing applicational solutions to us today.

A bot’s nature varies according to its developer’s skill. And by the developer we do not understand the software engineer, i.e. the one who programmed the software engine or the bot software shell, but basically the knowledge engineer, i.e. the author who creates the bot’s conversational scope, its recognition patterns and the range of knowledge topics as well.

Some further soft facts add up to this. Communicational results can only be quantified on the user’s side. If during a session the user acts according to a bot’s inherent communicational model, this bot has to be considered successful. Admittedly that is not quite true since the user may not know anything about optimising the internal communicational threads on the bot’s side. The Turing Test would not be meaningful in that respect since its test setting is solely designed to verify from the user’s point of view whether he or she is communicating to a bot.