Turing Test – A little less conversation

Our very own Elbot (www.elbot.com) was honoured last weekend when he was just one of five chatbots invited to take part in the latest ‘Turing Test’. It’s sixty years since the death of Alan Turing and rather than look at where artificial intelligence might be in another six decades we thought it might be more interesting to consider what is likely to happen in the next six years.

Until now, Elbot had been the closest to beating the Turing Test and we offer congratulations to Eugene for being the first one past the post. However, whilst the debate is still on as to whether the test itself is demanding enough, I think that deceiving a human into believing they are talking to a machine is not actually as important to the future of natural language as a machine that can speak, think and interact like a human.

There are several reasons behind this. Firstly screen real-estate on next generation devices such as wearables, is likely to be non-existent. Interaction is expected to be via speech and possible a degree of gesture, and this in turn will place emphasis not just on the quality of the speech technology, but also in the technology’s ability to reason and react like a human, thus demonstrating an underlying intelligence far beyond holding a conversation.

Implicit personalization, that is to say learning about a user, their likes and dislikes from their behaviour or conversation, is another area of NLI starting to take off. Initially, it will allow machines to give responses tailored to an individual, in the very near future it could see your personal assistant app offering to order from your favourite take-out every time you watch an action movie, simply because that’s what you always do.

Despite network operators’ best efforts to ensure internet access is everywhere we go, there are always times when connection is just not available. This will require for machines to still be able to interact with human in offline modes and to be able to store those requests that can’t be processed until it is back online.

In addition, organizations need to be able to learn from their users through analysis of conversations and behaviour in order to improve their own offerings. Sort of like a focus group, but instead of a group of twenty people, it’s scaled up to millions of inputs.

To achieve all this requires a degree of machine learning, which could be debated is nearer to what Alan Turing was trying to achieve when considering whether or not a computer can think for itself.

Don’t get me wrong, witty, interesting human like conversation is still important in the mix of human and machine interaction for many different reasons, but a positive user experience has to come first, so whether person thinks they are talking to a human or not isn’t necessarily the issue. To quote a line from a song made famous by Elvis, A little less conversation, a little more action please.


Andy, who lives with his family in the UK, is Chief Marketing & Strategy Officer at Artificial Solutions. A regular speaker at industry conferences and events, Andy delivers insight on the rise of AI, the challenges businesses face and the future of intelligent conversational applications.

Leave a Reply

Your email address will not be published. Required fields are marked *