Emerging Techs Multiple Paths

Over the next 10 to 15 years, the emerging technologies of today will completely reinvent entire industries. Not since the advent of the internet some 30 years ago will technology change how we work, how we move about, and how we store and share information. Though, if the quote “What is past is prologue” bears any truth, the government will largely miss out on these changes, many of which will be hugely beneficial for large enterprises like federal agencies.

To move past this, the government cannot assume that its security and data is too sensitive, that its processes are too unique, or simply rely on those managing the existing technology to determine what should be adopted next. The federal government has done quite a bit of work recently to demonstrate awareness and set policy for emerging technology, whether its autonomous vehicles, artificial intelligence (AI), or fintech. The limitation here is that these reports are almost always outward looking. Rarely do federal commissions, reports, or frameworks on subjects of emerging tech dive into how the government should study, pilot and adopt these technologies. But they should. This is especially true considering the federal government outlays account for about 20% of our GDP; would you take a government report on blockchain seriously that didn’t consider its implications on the healthcare industry?

“Innovation” is Different This Time

In the last 10 years, most governments have very cautiously adopted technologies like self-service cloud computing, internet-based communications, and big data processing, especially when compared to their commercial counterparts. While all of these aforementioned innovations spurred growth across our economy, they were really an improvement on existing technologies and processes. Technologies that are emerging today — such as machine learning, distributed ledgers, and AI — will be applied to all industries from automotive to healthcare and customer service. Unlike before, these technologies will drastically modify human behaviors and processes and will have massive effects on the larger economy and how we live and operate.

In order for the government to more rapidly and effectively adopt this emerging technology, each agency needs to make it a priority to research how it can use modern technology and apply it to the mission of the organization. In some cases, agencies will need to apply strategies that fall outside of the traditional processes of the organization. Agencies must at least ask themselves how well their problem is defined and who might be best suited to solve that problem, which in most cases will be a product or company sitting outside of government. Interestingly, AI and distributed databases (e.g. blockchain) may take different paths to adoption based on these considerations.

Above all, agencies need to have a product mindset when looking to adopt new technology. Organizations that look to buy a set of requirements will continue to operate in the same way they have in the past, whereas those that leverage commercially available solutions to help solve their problems and embrace new domains will be more likely to move along the adoption curve.

from Big Data to AI

In the last few years, artificial intelligene has overtaken big data as the “nom du jour” in emerging technology. This is unsurprising given that many of the problems big data initiatives can inadvertently cause (e.g. unmanageable data set sizes, noise in results and insights) can be improved through the use of AI. Agencies that have built or are building big data programs and leveraging big data technologies should absolutely be thinking about AI as a resource in their work. In fact, it is likely that agencies are already leveraging technology that rely on AI and don’t yet realize it. Google Analytics (the engine that runs analytics.usa.gov) released product functionality last year that relies on machine learning and predictive analytics.

A curious outcome from a recent survey of enterprises conducted by Narrative Science showed that while only 38% of enterprises said they were leveraging AI technologies, 88% of those same respondents had purchased solutions that relied on technologies that fit in this space (i.e. machine learning, natural language processing). If agencies can understand that leveraging these technologies organically will yield far less benefit than if there were a strategic approach, adoption would happen more rapidly and with more benefit.

Two areas that could be most useful for the government to explore in the AI space are predictive analytics and the automation of repetitive or manual tasks. These tools can be useful across dozens of mission areas from fraud identification and mitigation to more efficient resource allocation. Automation is a bit more complicated but possibly more immediately rewarding for an agency. Reallocating expensive human capital to more high value work can lead to massive cost savings but agencies must have a plan for how automation is going to affect the current workforce.

Beyond the reallocation of people resources, one of the biggest barriers to AI adoption in federal agencies at this point in time is the lack of data science skillsets across government and the dozens, if not hundreds, of different tools that are being used for data analysis today. Three Dcode42 portfolio companies that are making this a more manageable challenge are DataRobot, Domino Data Lab, and Catalytic. DataRobot has an incredible product that democratizes data science and makes it accessible to analysts without a data science background. This can allow agencies to see the power of AI in their data without waiting for skills to catch up with technology. Domino Data Lab enables organizations with different tools, code bases, and skillsets in data science to more seamlessly experiment and drive innovation into their organizations without having to worry about legacy technology issues or siloed efforts. Catalytic allows agencies to quickly apply AI to the underpinning operational processes without huge integration efforts — drastically reducing the need for human involvement in many cases. The main point here is that products are available to not only begin leveraging technology, but also overcome adoption issues as well.

With AI, many agencies are already moving in the right direction. But there is more agencies can do to understand the value of this technology. Starting with small but targeted pilots is always a good idea and, in some cases, creating a semi-autonomous innovation team can be helpful in spurring new ideas and adoption of new technologies. AI technology is a natural extension of the advent of big data and agencies need to apply a strategy to its adoption.

Embracing a loss of control

While big data in government may have prepared agencies for the arrival of machine learning, blockchain technology is going to be a larger chasm to cross. This is primarily because government operations often resemble a traditional clearing house, which is exactly what distributed ledger technology was built to disrupt. However, the implications for the use of blockchain in government is massive. Improper payments is a $125+ billion problem in the executive branch, the access to and security of personal data is woeful in the federal government and the $500 billion procurement marketplace is full of inefficient chokepoints and supply chain risk. All three of these areas and many others could be improved by blockchain and there should be active pilots ongoing across government.

In the case of blockchain adoption, we suggest that agencies should carve out groups to examine this work separate from the larger organization. In many cases, the authority derived from approving, clearing, or aggregating information and transactions will be taken away from the central owner of that process. This is a poor place to put a disruptive innovation. This is not because those organizations are “bad” at innovation. In fact, in some cases centrally housing work is a good idea. In this case, government shouldn’t expect a disruptive technology to be incubated by the organization which may derive value from the existing process. Clayton Christensen and Michael Oversdorf prescribe creating “heavyweight teams” to tackle this type of disruptive innovation. Government should heed this advice and build teams to understand blockchain, how it can be applied to known problems, and where the private sector is moving.

It has been almost three years since Microsoft debuted “Blockchain-as-a-Service” and private sector consortiums are consolidating around blockchain technologies, whether cryptocurrencies, platforms or tools. As of February, the Government of Georgia is now using Bitcoin to validate property level transactions. This was a first for any national government. Many U.S. states are using legislation to create entities to explore use around smart contracts, voting, gun ownership and registration, and for use in tourism. And there is even a Congressional Blockchain Caucus. Beyond regulating the market, the executive branch needs to step out and understand this technology. A single agency can spur others to act, and while DOD is working to understand this technology through DARPA, a civilian agency would be welcomed as a first mover as well.

While it isn’t precisely clear when and how these two emerging technologies will affect our economy and lives, the likelihood of disruptive change happening in some form is far greater than it is not. The federal government needs to be ready to apply the correct models of experimentation and adoption to both of these technologies and begin testing the use of certain products against its problems and processes.

This story was cross published here and here.