(Image by Nuance)
If you’re fairly new to the world of technology, you many not have heard of the trend to ‘democratize’ data – a trend that arguably began in 2013. The Innovation Enterprise describes how data has changed from being held and understood by a handful of highly paid professionals to being accessible to all levels of employees and the general public at large. They refer to this process as data democratization. One of the assumptions behind democratization is that the crowd is wiser than the individual and that great innovations can happen when everyone has access to the data.
If you recall from one of my earlier blogs, I talk about the role that machine learning plays in the field of data science. Basically machine learning is a way to find patterns from the past in order to predict what can possibly happen in the future. Today I want to dive deeper into a December 2016 Forbes article that presented the argument for democratizing machine learning. In the article, Francois Chollet, a Google researcher says that it’s important to democratize machine learning for two reasons: to not let anyone’s contributions go to waste and for social and economic stability.
The first reason that more people should have access to machine learning knowledge and their tools is so the environment is rich for people to be inspired to create new innovations. Democratizing machine learning is a key strategy to have interested and talented people add value to existing products. In Jared Dean’s book Big Data, Data Mining, and Machine Learning: Value Creation for Business, he underscores the importance of putting the data and tools in the hands of lots of folks in order to spark creative genius. And as Atul Butte from Stanford University aptly summarizes, ‘Hiding within those mounds of data is knowledge that could change the life of a patient, or change the world.’
In addition to helping drive innovation on a larger scale and more quickly, Chollet thinks that machine learning and artificial intelligence will automate many jobs in the future. By allowing the crowd to generate value, the loss of these jobs will be balanced out since individuals will have power over the machine learning technology. It’s interesting to think you could develop a automation tool that would eliminate your job and that somehow you’d benefit from essentially making yourself obsolete. I’m not quite sure most people would be enthusiastic to create if they knew their creation could possibly destroy their livelihood. However tenuous an argument, democratizing machine learning is happening and will continue to probably grow to some degree in the future.