The AI revolution will never happen in the humanitarian sector, it’s already here.
When I started Relief Applications back in 2013, we were all speaking about Mobile Data Collection. I was deployed during the Haiyan cyclone with the Red Cross and we put a lot of efforts setting up a good old Open Data Kit questionnaire.
Last month I was making a presentation to the International Labour Organisation in Geneva focused on new technologies and future trends and was insisting on the fact that AI, and especially Machine Learning, is integral to the current work and operations of humanitarian and development organizations. It is not a distant future technology but today’s reality.
So what happened? How come AI is everywhere now and we did not even see it coming?
I am not going to go through go back on the whole history of AI. In my generation AI was just a researcher topic with some funny breakthrough like Deep Blue beating the work chess champion Kasparov back in 97 or AlphaGo accumulating victories in Go since 2015. But without realizing, all major IT companies were developing concrete applications for the general public. Apple launched Siri in 2011, Watson by IBM in 2011, Alexa by Amazon in 2014 and the list goes on. These easy to use applications (packed with incredible tech behind the scene) are becoming part of our everyday life discreetly entering our work activities.
Consuming AI APIs (Application Programming Interface) or simply put using AI algorithms is a matter of few lines of very simple code. In some cases, modern AI interfaces allow the use of advanced algorithms with some clicks on a well designed web interface. Look for example at the Google AutoML project where you literally just have to upload images from your phone and tag them to create a Machine Learning Algorithm that will recognize pictures for you in a matter of few clicks. Take few hundreds pictures of your favorite food, let’s say croissant, upload them, wait few minutes for your AI algorithm to be ready and present any picture to your algorithm and it will tell you if it is a croissant or not! Or the Amazon Alexa project were the Amazon algorithm learns to recognize the questions you ask to give you more and more accurate answers.
In our sector, there have always been two worlds coexisting; people in the field running programs with extremely low tech systems and IT experts and university researchers developing incredible pilot projects that have extreme difficulties to be replicated as they rely on very, very high tech that only a hand full a people are able to comprehend. This rule applies to AI, there are some incredible research field to be explored but very little is done at field level, for concrete and hands on applications.
But the real transformation is coming from every day AI, bots that conduct surveys for us, smart email categorization, natural language analysis, smart personal assistant, smart data search, voice-to-text and text-to-voice algorithm, you name it.
The AI revolution is never going to happen in the humanitarian sector because AI is already here. We have, as humanitarian actors, to shift our mindset and leverage AI capabilities into our daily work. It is a constant battle for organization to think innovation and millions are spent on innovation teams but to some extend the real change will not come from these nice and wishful pilot projects but from our capacity to use already existing AI applications in our daily jobs.