Machine Learning Chatbots Create Their Own Language
by Jon Wood
I have always thought that chatbots were a bit of a gimmick. I think they’re a smart idea and they are quite funny and responsive, but up until now, I have always seen them as a bit of fun. However, after reading about what’s just happened in Facebook’s artificial intelligence lab, my eyes have been opened.
For the uninitiated, a brief explanation: a chatbot (also known as a talkbot, chatterbot, Bot, chatterbox, Artificial Conversational Entity) is a computer program which conducts a conversation via auditory or textual methods.
Some chatbots use sophisticated natural language processing systems, but many simpler systems scan for keywords within the input, then pull a reply with the most matching keywords, or the most similar wording pattern, from a database
Chatbots are used in call centres all around the world; we speak and negotiate with them every day. Yes, it’s all in a very basic format for now, but they do have a purpose in the consumer space which is already being taken advantage of. It’s just what’s going on in Silicon Valley’s labs that’s taking these bots to the next level…
What Have These Machine Learning Chatbots Been Up To, Then?
Researchers at Facebook have found a buried line in their latest AI report. It seems that their chatbots are having conversations with one another, thus providing a fantastic glimpse into the future and, in particular, the future of language.
They were even able to negotiate with each other. It’s hard to imagine chatbots negotiating but this really could be the future. This means that there is the potential necessity for us humans to learn their language; the language of bots (Botanese?).
Researchers at the Facebook Artificial Intelligence Research Lab describe using machine learning to negotiate and to sell – it turns out bots are actually quite good at deal making! Albeit by diverging from human language in order to do so.
In layman terms, the model that allowed two bots to have a conversation—and use machine learning to constantly iterate strategies for that conversation along the way—led to those bots communicating in their own non-human language that the researchers are not yet able to understand. And because of the power of machine learning, it’s probable that their language and capabilities will be able to evolve fast – will we be able to keep up? And what happens if we can’t?
All this news about these incredible machine learning chatbots really did fill me with disbelief and a real sense of wonder. We are just beginning to understand the future of relationships between machines and humanity, and it sounds pretty crazy.
A lot of the artificial intelligence that is being developed at present uses machine learning algorithms modelled on the neural networks of the human brain, hence why the term ‘artificial neural networks’ comes up so much in articles and papers on AI. So it shouldn’t come as a very big surprise that, as these systems become more sophisticated, they begin to mimic human behaviours that we have always thought made us so special.
Don’t worry though, the Singularity isn’t here…yet. However, what Facebook has done is to demonstrate how these machine learning chatbots, as well as all the other AI tech coming up through the ranks, are redefining people’s understanding of so many realms once believed to be exclusively human—like language. It certainly is an exciting and awe-inspiring time to be living in; a time when we really begin to redefine what it means to be human, and equally to begin to ask questions about what to do with the Frankenstein’s monsters we are creating…