The Future of Conversational AI

Grace Hughes and Rui Teimao are both experts in conversational design, with the former based in Dublin and the latter in the UK. Both have a passion in exploring the importance of words in technology and design — and wanted to share their insights with Rogue Bees.

As AI technology advances, the need for conversational design becomes paramount. Designers have always been interested in language — both in words and the shape of online spaces. As part of this, both Grace and Rui are part of the Conversation Design Hub, building a community that explores the intersection of technology and language and formalizing a framework for conversation design.

Grace and Rui found that, in the world of chatbots and conversational interfaces, the two main types of conversations are transactional and social:

• Transactional conversations, which are more popularized, are those we have with chatbots to solve a problem or obtain information quickly.
• Social conversations, on the other hand, are more about building a relationship or providing emotional support.

For example, a chatbot designed to provide mental health support may ask questions such as “How are you feeling today?” or “How was your day?” to establish rapport with the user. Financial services call for transactional conversations, getting things done quickly and efficiently — however, the need for social or more ‘empathetic’ conversations may arise when dealing with more vulnerable users like those in poverty or others who may be grieving and dealing with the financial affairs of a loved one who has passed away. Also, the more frequent use of a chatbot or a voice assistant may shift the popularity of AI interactions from transactional to social conversations.

Additionally, both emphasized the importance of understanding the different types of conversations that can take place through chatbots and conversational interfaces. While transactional conversations are straightforward, social conversations require a deeper understanding of human emotions and communication nuances. As chatbots become more ubiquitous, it is essential to consider the ethical implications of these interactions and ensure that users are not put at risk.

That is why conversations take into account how humans can be more direct with AI. When we think we’re talking to a human, we tend to talk too much, interrupt, digress, and do things that the chatbot can’t handle. But when we’re transparent and know that we’re talking to a system or software, we tend to modify our language and use very short, direct sentences. This directness helps the channel convey the content.

Other benefits exist, too. Using natural language and conversational techniques can make interactions feel more comfortable and familiar, potentially creating a deeper connection between the user and the machine. This is where personality comes in — even a neutral personality can help shape the conversational experience and create a more cohesive system.

Another important consideration is what happens when things go awry. With conversational interfaces, this is bound to happen frequently. While conversation can be designed and planned, humans are unpredictable, and language is complicated. This presents a major challenge for AI systems, as they must be able to adapt and understand what people mean, even when language is interpreted differently.

One major example Rui outlines is for people with low literacy. GPT models are effective at transforming text from one level of literacy to another, which can bring more clarity to topics. They can rewrite text for a seven-year-old or with greater detail for a post-grad student. Applying this technology to filter inputs and outputs can eliminate literacy, language, and country differences. For example, a project aimed to provide multilingual customer support using pre-written English scripts for agents in India. Customers from various countries can call in and ask questions in their language, which is then translated and the appropriate response is given in English, which is then translated back to the customer’s language. This technology has the potential to allow anyone in the world to communicate with anyone else in their own language, breaking down language barriers and bringing people together.

What about human creativity? In short, Grace and Rui are not worried about its impact on human creativity. Both see AI as a powerful tool that won’t replace the human spark of creativity and playfulness that is unique to humans but rather act as a teammate.

As AI continues to evolve, it is vital to recognize its strengths and limitations. It is an incredible aid that can accelerate and automate many tasks, but it lacks the human touch — creativity and imagination.