Alana is a business driven by research, which means we’re inherently predisposed to look ahead. So, here we invite you to take a peek into our minds as we share our top 10 predictions for what’s likely to preoccupy the world of conversational AI in 2021.
Conversational AI, in all its guises, is changing the way we interact. As it expands into more and more aspects of our everyday life, we’re happily revealing more to the AI assistants in our homes, our offices and in public environments. Advancing security to guarantee conversations remain private and user data is secure will continue to be a key priority in 2021.
Ethical conversation design
How accurately can AI assistants identify and filter out offensive content before including it in answers? How can we prevent home assistants unintentionally sharing fake news and misinformation? Do we need to be more mindful of how gender, racial and other bias could creep into how we’re training machines? As AI conversations become more wide-ranging, we need to build a finely-tuned ethical conscience into technology.
Natural conversations are often sparked by something we can see. And we instinctively use imagery to explain concepts and our needs. As we seek to make AI conversations more human-like, the ability to respond to visual cues and incorporate image references into AI voice conversations will be an exciting progression.
Voice search is now a rapidly growing form of access to information, but to be even more useful, it will need to become more conversational. Multiple conversational turns, follow-up on search responses, clarification and refining searches – are all aspects of natural conversation that Conversational AI like Alana, is beginning to replicate.
Interacting in AR and VR
Today, you can put on a set of VR goggles to immerse yourself in gaming or go snorkelling on the Great Barrier Reef. It’s an immersive visual experience, but what about conversation? Imagine if you could use conversational AI to talk to other people about places and objects in a graphical or augmented reality environment.
We’ve been tackling an unexpected spike in isolation and loneliness in 2020 as we live under the restrictions inflicted by a global pandemic. The demand for AI assistants to fulfil a companion role is something we’re expecting to grow in 2021.
Human conversations are rapid exchanges, sometimes with minimal or no pauses between speech turns. Conversations with AI devices tend to be more stilted and lack natural fluency. Reducing latency and working towards incremental, word-by-word processing of conversations continues to be an important ambition for conversational AI going into 2021.
To engage in a more natural interaction with a human, robots need to be able to figure out what you can see and react verbally to the environment surrounding your conversation. The potential to give smart assistants a physical form is an exciting direction for the future of AI conversations.
Doing more with less (data)
The notorious GPT-3 language generator has 175 billion parameters. It reportedly cost $4.6 million to train and has a huge carbon footprint. Very few organisations have access to the huge amounts of data needed for models like this. So the focus shifts in 2021 to how to build and train exceptional language-generating AI machines that use less energy and require much smaller amounts of data.
Alana, can you hoover the lounge while I’m out?
Are our AI home assistants one step away from becoming physical, robotic companions that can make small talk as they unpack your groceries? Not yet. But with the use of voice assistants expected to triple over the next few years, we can certainly expect to hold more useful and entertaining AI conversations in the year ahead.