Short Bytes: The proliferation of new applications for machine learning hints at its vast potential. Companies, governments, and researchers innovate new ways to mine big data for useful insights constantly. All it takes is some creativity and computing resources to turn data into intelligence. Read more to get an insight about the future business implications of machine learning.
Having good access to information and being able to spread information quickly has been a hallmark of great revolutions in technology, from Gutenberg’s first printing press to the rise of the telegraph during the American Civil War, Bell’s work with the telephone, and, of course, the Internet. Both communications and data collection are fundamental not only for business decisions but also for medicine, governance, and many other fields. Big data is a quantitative evolution of the way people can accumulate information, but the sheer scale of large datasets yields an unprecedented level of detail for many applications. The difficulty of creating and manipulating a large dataset is dropping by the month, with simple Hadoop reporting taking the place of cumbersome SAS jobs and other outdated tools.
Consider auto insurance. Right now, auto insurance carriers use complex statistical models that use demographic information about their customers to predict which of them are most likely to get into an accident and make a claim. They need to pay actuaries large amounts of money to build and test these models, and then they base their pricing on the models’ results. Recently, however, some insurers have decided to start using actual driving habits to price their products. They ask customers to install a device that plugs into their car and records speed, how fast the driver brakes, how tight they turn, and other metrics. Then they use that data to predict accidents over time as well as use this information to generate a more personalized auto insurance quotes comparison. This is a striking and creative use of data to solve a business problem: pricing, but in the near future it could go even further. Fine enough data would let the device predict when a driver was intoxicated, tired, or distracted based on their driving actions. This kind of data could save lives.
Data collection about patient health records could be subjected to machine learning analysis, helping doctors predict disease and illness with better certainty than they can now. On the dark side of big data, governments can train machine learning programs on datasets containing the personal information of their citizens, making it easier to monitor their daily lives. It is a fine balance between using machine learning and big data to improve people’s lives and violating privacy on a massive scale.
Right now, it comes down to case by case judgments. The legal system is not well equipped to handle questions pertaining to the details of statistics and technology, but that will need to change. Big data is not as flashy as some previous technological revolutions because it takes place in the silent realm of software, and machine learning is a much more obtuse concept than the telephone or the printed book. However, its potential to transform society is just as large, and big data stands poised to alter the way we think about privacy, prediction, business, and relationships.
The proliferation of new applications for machine learning hints at this potential. Companies, governments, and researchers innovate new ways to mine big data for useful insights constantly. All it takes is some creativity and computing resources to turn data into intelligence. Machine learning threatens to alter entire fields, like medicine and law, and to forever change the way we think about information. Soon, gleaning insight from data won’t be the exclusive realm of humans, but of algorithms. The task that remains in the hands of people is to decide what to do with that insight.
Share your views in the comments below.