Can We Trust Chatbots?
A lot of people don’t understand how technology works. Technological development has seemingly exploded at a rapid rate in just one generation, and with anything if you do not choose to study it or engage in a subject it seems inaccessible and strange. Complicating matters is the media, in particular film and TV’s instance on portraying technological advancements as often dangerous, be it killer robots or technology taking control of your environment.
Yet there is so much potential for good in technology, and so many people are advancing the world with it. Whether it is mechanical advancements for transport, new medical tools or ways to combat domestic violence.
Enter chatbots. Some people are wary of a machine talking to them or giving them information. They can be suspicious, asking questions such as where does the information come from or who is speaking? But AI chatbots are not the future, they are the present and increasingly found to be useful by different businesses.
Conversational AIs and chatbots will have different degrees of intelligence and names but many can also learn and become better conversationalists via machine learning. You will not feel like you are talking to a robot and instead they are increasingly designed for trust, seeming friendly, listening and being transparent through natural conversations.
They offer many benefits to companies and users. Firstly, instead of a user having to wait for a staff member to be free, a chatbot can provide instant responses 24/7, everyday of the year, to multiple users simultaneously and in multiple languages. This broadens the spectrum of who can access information and get help, as well as the rate at which this can be done. Opposed to one on one staff to client ratios, you can have thousand to one client to chatbot ratios. They also reduce human error, an AI doesn’t forget things like a human would. This reduces the burden on human agents, and enables an increased scale of reach for the business.
AIs and ethics are a huge topic, but we will briefly touch on two major ones for chatbots: data biases and data collection. Whilst there can be issues with data input, e.g. biases encoded by a team based on their understanding of and movement in the world such as through gender, ethnicity and language, increasing diversity within a team is beneficial to reduce this and not replicate biases inherent in society. A chatbot can also be trained through extensive data, and human intervention can continually finetune it. This means that AIs can constantly develop and become better. In regard to data collection and privacy many chatbots are transparent about who the user is talking to and what data, if any, is being stored and/or removed, as recommended by UNESCO’s ‘do no harm’ recommendations. Many minimize data collection, which makes users feel safer to use AIs.
Here at We Encourage we have created our AINO chatbot that is learning how to respond to people who may be experiencing domestic violence, and to provide information. We have trained its current responses but we are building a more advanced version so it can become conversational. For us, a non-judgemental digital service that collects no personal data and can be used quickly, securely and silently in the potentially dangerous situation of domestic violence is a great way to utilize chatbot technology. Please access it if you have questions about domestic violence, or your own situation, and share it to others to help them and help AINO become better at its job.
We hope that this has been helpful for you, and if you are looking for more information check out our website or our AINO Chatbot.
