The gold mine hidden in government’s data dustbin
In 2011, McKinsey did a study estimating just how transformative the implementation of big data and artificial intelligence could be for the public sector in the European Union.
Even back then, McKinsey estimated potential savings to be approximately $300 billion annually. Furthermore, the consultancy concluded a similar savings would be likely if the technology were applied to the U.S. public sector as well.
Savings of $300 billion is the equivalent of a 40% reduction in the current annual Defense Department budget of $750 billion. Let that sink in.
McKinsey also found that a paltry 20% of those savings are being realized in the United States.
So far, most of these savings come from leveraging structured data to make government services more efficient by reducing fraud and errors in transfer payments and improving tax collection.
With the volume of structured and unstructured data doubling every three years in the U.S., much of it is being swept into the government’s data dustbin because leveraging unstructured data has proved to be too difficult.
Interestingly, most of the potential to improve access to unstructured data exists in automating analysis of large collections of audio, images and video.
Historically, pulling insights from unstructured data has required assigning people to these problems. For example, many Defense Department analysts are tasked with reviewing video feeds and painstakingly summarizing what they see every day. The military refers to these manual problems as cognitive loads and sees human-AI teaming as a way to solve them. In other words, offsetting rote mechanical processes onto machines and algorithms frees up mental cycles for human beings.
Knowing this, why has progress been so slow when the need is so great and the value so readily apparent? Three major structural reasons:
- Legacy technology issues.
- Siloed data and agencies’ reluctance to share it.
- Scarce talent resulting from government salaries and hiring requirements, coupled with the huge competing private-sector demand.
Fortunately, these are all solvable problems.
The resolution of the first two requires government agencies to recognize that the inability to tap into unstructured data is a major national security issue from both a military and economic perspective. Competing foreign governments, such as China, are not wasting time as they digitally transform their own public sectors, so the U.S. must move quickly to remain competitive and secure globally.
To remain competitive, talent is the ultimate key. Hiring the best people to do the work of streamlining and automating public-sector processes requires the government to pay competitive salaries for engineers and data scientists. The formation of the U.S. Digital Corps is a step in the right direction, as it seeks to recruit the next generation of technology talent to federal service, but Congress must address the talent pay gap head on.
Meanwhile, other solutions must be explored. Commercial technologies such as robotic process automation may plug some critical gaps. By leveraging AI to streamline government operations, these technologies both modernize agencies and buy them time to recruit the technical talent of the future, enabling America to remain competitive on the global scale.
Digital transformation is not simply a matter of implementing new technologies and systems. It’s also a matter of aligning them to fit the 21st century. Government agencies must become less compartmentalized and more willing to evolve.
It’s clear that adopting a hybrid strategy of modern technologies such as AI, while structurally reworking how organizations operate, will make the U.S. government more responsive and efficient and, ultimately, make us all more prosperous and secure.
Dan Morozoff is co-founder VIDROVR, a machine-learning company.