Michael McTear is an Emeritus Professor at Ulster University. He has been researching in the sector of spoken dialogue systems for greater than 20 years and is the writer of several books, including most recently Conversational AI: Dialogue Systems, Conversational Agents, and Chatbots (Springer Link 2021). Michael has delivered keynote addresses and tutorials at many academic conferences and workshops. Currently Michael is involved in several research and development projects investigating using conversational agents in mental health support and the house monitoring of older adults.

What initially attracted you to machine learning?

Until recently I even have worked with rule-based approaches to conversational AI, particularly in the world of dialogue management where the essential idea is to develop rules that determine the agent’s strategies in a dialogue. Nonetheless, with recent advances in machine learning, first in reinforcement learning after which in deep learning, I even have found that these approaches can potentially address a few of the issues faced by rule-based methods, corresponding to the issue of scaling and the necessity to write down multiple rules to cater for more complex dialogue flows.

You’ve been working on voice and conversational AI for over 20 years, what made you give attention to this field?

I even have been desirous about how conversation works for a very long time. In my PhD thesis I studied the event of conversational competence in young children and this was the subject of my first book. Later I became intrigued by the concept that computers could engage in conversations in a human-like way and since then I even have followed developments on this area. At first it was very much a minor area inside AI but throughout the past few years it has change into very essential because it has been adopted by the massive tech corporations as an area of strategic interest for them.

One in all your most up-to-date projects that you simply focused on is ChatPAL to advertise mental well-being in rural areas. Could you discuss the challenges behind constructing a chatbot for users who will not be tech savvy or acquainted with the concept of chatbots?

Many individuals are acquainted with voice assistants on their smartphones and as smart speakers corresponding to Amazon Alexa. Young people use their phones mainly to text and in order that they are acquainted with the concept of interacting with a text-based chatbot. Nonetheless, in relation to interacting with a chatbot that’s specialised for a selected domain, as in our Chatpal project that was developed to advertise mental well-being in rural areas, we found that some users had expectations concerning the technology based on their experiences with Alexa and other similar chatbots that exceeded what we were in a position to offer with more limited resources. We attempted to deal with the difficulty of users who weren’t tech savvy or acquainted with chatbots through initial living lab sessions in addition to ensuring that interactions with the chatbot were natural and intuitive.

What are a few of the challenges behind constructing chatbots which can be focused on mental health?

There’s the danger that some users might expect more from the chatbot that is feasible with current technology. We didn’t wish to change into involved in doing any diagnosis as we felt that this is simply too dangerous and there have been reports of chatbot responses in such situations that could possibly be considered harmful and even dangerous. We were guided by the necessities of assorted ethics committees in addition to recognised standards for the design and development of chatbots. One other issue is that we found that there have been differences between users when it comes to how they used the chatbot. Some users gave up quickly after they experienced technical issues whereas others were prepared to persist. There also was an age-related difference as younger users were joyful to interact with our text-based chatbot whereas older users were less joyful with this type of interface.

Among the apps that you simply’ve worked on offer motion plans for users, how do you effectively generate user motivation in an app?

To do that it’s essential to create and maintain profiles for every user that reflect things like their upcoming appointments, medications, general preferences, and what they’ve discussed in previous conversations with the chatbot. Users often state that they need the chatbot to pay attention to their individual needs and to maintain track of what has been discussed before quite than providing a more generic and fewer adaptive interaction.

Nonetheless, against this, there are issues of information privacy and users also express a priority concerning the use of their private information. There’s a fragile balance to be struck here and naturally there may be an ever increasing amount of laws to manage the moral use of AI in private and non-private life.

What are some ethical considerations behind constructing chatbots?

One in all the principal ethical considerations to think about  when constructing chatbots is whether or not they reinforce gender stereotypes. Traditionally females have taken assistant-type roles within the workplace while males have assumed leadership roles. Implementing a chatbot with a female persona could reinforce such gender stereotypes.

One other essential ethical issue is whether or not chatbots should embrace human values and behave in a way that inspires trust. Thus is referred to as the alignment problem. Chatbots must be  designed in order that they avoid breaching human rights and creating bias, and their decisions must be transparent to human users.

Also, as mentioned earlier, chatbots should respect user privacy and data protection laws. There’s loads of research and energy being devoted currently to those ethical considerations.

In a world that is targeted on English speaking chatbots, what are some challenges behind designing multi-lingual and international chatbots?

All of it will depend on the provision of language resources corresponding to language models and, for voice-based systems, speech recognition and speech synthesis engines. This is just not an issue for widely spoken languages but difficult for languages with limited resources which will nevertheless be spoken by a lot of people and where there may be a definite need for the services of a chatbot. One possible solution is to make use of transfer learning from a model pretrained in a language corresponding to English and wonderful tune it with data from the low resource language.

Many of the apps that you simply’ve designed use open source software, what are a few of the perfect open source tools on the market?

Using open source software was a requirement of the agencies that funded our projects.

We’ve got used Rasa in our projects because it is open source but additionally very powerful because it makes use of the newest developments in conversational AI technologies. In addition to Rasa there are several excellent open source conversational AI software products, including: Botpress, Microsoft Bot Framework, OpenDialog, and DeepPavlov, to say just a number of.

You’ll be speaking on the upcoming Way forward for Chatbots & Conversational AI Summit, what is going to you be discussing?

In my talk I will likely be comparing the standard approaches to chatbot development based on best practices, referred to as conversation design, with latest approaches based on large language models corresponding to ChatGPT. I’ll cover the professionals and cons of every approach and argue that, although approaches based on large language models offer loads of potential for the long run development of chatbots, there are still many issues with the uncontrolled use of enormous language models, especially in areas corresponding to healthcare and business, in order that there remains to be a necessity for conversation designers who can ensure explainable, transparent, and ethical conversational AI.