Hallucination: AI is at its best when it mimics the worst human behaviour

I like the fact computers behave unpredictably. Given an identical instruction on two different occasions, a computer will do something different each time. Conversely, humans are totally predictable, and make the same dumb mistakes over and over. Until now, when I’ve pointed this out, people laugh as if it’s meant to be a joke. But is it? And how is it relevant to research?

There’s been a lot of attention paid to the use of artificial intelligence in Chatbots following the release of ChatGPT. Chatbots are used by businesses and organisations to listen to and respond to customers. For customers, they are an alternative to navigating endless options in an application or on a website. Chatbots replace human interaction, thereby saving costs. Businesses and organisations that use these chatbots buy the module from a chatbot supplier. Likewise, market and social researchers are looking at Chatbots.

Off-the-shelf chatbots are initially void of information. The purchaser must feed the necessary information into the chatbot before it will work.