The United States government is the single largest buyer in the world, spending over $500 billion each year, with a substantial percentage of that spend allocated to IT procurement. However, despite spending large sums of money procuring and sustaining IT, the U.S. government lags behind commercial organizations in terms of technical maturity.
This isn’t for lack of trying; there are countless logistical and bureaucratic hurdles to governmental technical maturation. Between complex security requirements and institutionalized bureaucracy that makes rapid IT procurement challenging, it’s difficult to evolve a federal tech stack. The U.S. Department of Defense (DoD) reported that developing an Artificial Intelligence (AI)-enabled Minimum Viable Product typically takes corporate technology companies six to nine months, but takes DoD an average of 91 months.
But the bigger issue is that the government is averse to the risk that a new solution could cost time and money to procure and then ultimately be less effective than an (inefficient) current solution. While the Silicon Valley mentality is to fail often and quickly as a means to more rapid learning, and ultimately, progress, that’s not always a viable option in defense, where mistakes may mean soldier or civilian casualties.
Steps Towards Maturity
The goal of any defense organization is to deliver its mission as efficiently and as effectively as possible. And in the long run, evolution doesn’t need to mean failure; by using a strategic iterative approach, government organizations can take advantage of more powerful technologies while mitigating the risk of missteps.
Ultimately, there is significant risk in the delayed adoption of powerful data-driven tools, especially in light of the pace at which adversaries innovate; responsible operations require government organizations to operate with as much precision and efficiency as possible, and AI equips them with the tools to do just that.
Not all AI is created equal, but there are a few best practices government organizations can adopt to improve their strategic AI iteration and ensure technological advantages over adversaries:
- Leveraging Probabilistic Models. Deterministic decision models based on static rules (e.g., If X, then Y) are comforting in their simplicity, but lag far behind sophisticated AI and machine learning, which can weigh a host of variables and concerns in order to make recommendations. Probabilistic modeling and partial dependency plots provide weighted recommendations and demonstrate the extend to which specific variables contributed to that recommendation, providing transparency into elusive mathematical models.
- Embracing Open Source. While privacy is paramount when it comes to government data and systems, open source technologies offer exceptional AI technology for free. The collaborative development community and widespread buy-in leads to software that performs better and is more reliable than exclusively private alternatives. The most innovative tools when it comes to building and applying machine learning are open source, but when it comes to operationalizing projects, organizations often need another proprietary layer to best integrate with their use cases.
- Investing in Tools that Empower.The right data tools demystify AI and enable everyone to understand how AI systems arrive at their recommendations and why. Additionally, a central environment with auditing capacity offers the collaborative experience that data analytics and decision-making flourish in, while restricting user access to sensitive data. This enables engagement without compliance risk, with the records to support any actions in the event of an audit.