Communication is the most important skill we possess as a species and it is about to be radically improved by the use of artificial intelligence.
One of our earliest taught skills from birth, verbal communication, allows our infant self to tell the world what we need, want and desire. As we get older our vocabulary increases to be able to describe more of this world and at the same time, communication becomes two-way between giver and receiver rather than our original one-way form of communicating.
We age, our communication further develops and our communication style differs depending on who we are talking with, i.e., we communicate differently with family, friends, co-workers and strangers. The setting for communication matters a great deal.
Communication is also strongly linked to our social evolution as a species throughout history. As our tasks and activities become more sophisticated, the words needed to carry out those tasks also become more sophisticated. A remote tribe, for example, uses a language and communication style that fits its day-to-day tasks and way of life. The same is true for those that live in a modern, metropolitan city. Linguistic sophistication, therefore, evolves immediately in those moments when we need to perform more complex tasks. In this way, language mirrors societies’ needs and uses over time and by necessity.
How we communicate is also complex, layered in meaning and open to wide interpretation. This is why it is so fantastic when it works and so frustrating when it doesn’t. It has slowly evolved from the earliest grunts and cave paintings to our modern set of global languages both verbal and written. Civilizations throughout history have invented new languages and adapted existing ones to create better communication.
Our modern language is now at a point in time where its evolution is directly impacted by technology. In the 20th century, this has largely taken the form of new hardware from the telegram and telephone to today’s smartphone and laptop. The next stage in evolution is coming from machine intervention in the form of various artificial intelligence. So, how can we directly benefit from this machine intervention approach?
Let’s consider what happens when I write an email. There are many aspects of the email which I must consider in my email – word usage, urgency, tonality, sentence structure, language, timing of the message, etc. As the writer of the email, I must make a decision on every one of these elements. All of these elements, however, can be improved upon using data and the application of machine learning.
To compose an email using machine learning each element of the email needs to be reduced to its component parts. This allows information to be collected and observations to be made. The idea is to have an idea of the meaning of the message in context.
“Fire Sale – Today Only!”
“Fire Alarm test at 9am.”
“The building is on Fire.”
Looking at each of these email subject headings we immediately know the importance and urgency of each. Without further information, we can make decisions based on this one-way communication. A machine reading of the message needs to understand the context and meaning of the words, then take an appropriate action on our behalf. This becomes one-way machine learning can be used to augment and assist in a task.
For a more complex example where there will be two-way communication, let’s consider the timing element of the message. The most usual time to send an email is as soon as it has been written. This assumes that the time the message is sent is the optimal time for the message to be read. However, this is almost never the case! Consciously or not, we all respond differently to email messages depending on the time of day – some of us read emails immediately, some save them for later with earlier emails getting buried beneath more and more incoming emails.
Assume for a moment that your email message is being sent to a work colleague, who is a similar rank to you, in the same time zone, and that the subject is not time sensitive or urgent. To use machine learning and optimize communication, we could look at all the times of the day that your colleague actively checks and reads email. Is there a pattern that can be used for further investigation? Data on the timing of active email use, email senders, and email subjects can all be used to construct a view of your colleague’s routines and optimize how and when your email message is received. Taking this ‘single-task artificial intelligence approach’, both machine learning and a knowledge base of relevant data can be used to carry out a single task, for instance, better timing. The end result is that an element is completed with greater efficiency and is optimized for greater effectiveness.
What should be taken from the example are the great number of variables and considerations that need to be taken into account for any single communication instance.
At the heart of good human-to-human communication is the need to make a message well received and understood. Where there is machine assistance, the effect needs to be invisible. Any hint that a message has been manipulated or altered using external data will be seen as disingenuous with all of the negative connotations attached.
What artificial intelligence can’t do successfully at the moment is replace the human altogether, in every situation. Keeping in mind that the purpose of communication is still to make a need, want or desire understood. These very human characteristics are the thing that sets us apart from machines.
AI, done well, will greatly improve every aspect of how we communicate with each other and will hardly be seen. This use of data will help bring us closer together while at the same time celebrating our individuality.
Read more about Knowmail AI for email – Try it now!