Page 20 - Federal Computer Week, March/April 2019

Artificial Intelligence
Integrating AI into
mission goals
Successful strategies focus on agencies’ broader objectives and invite employee participation
Josh Sullivan
Senior Vice President of Analytics, Booz Allen Hamilton
a realistic understanding of what the technologies can and cannot do.
Second, organizations need to assess whether they have the right capacity in terms of talent, infrastructure, data, policies and culture to support and maintain AI solutions. Performing quick assessments of each business line against an AI readiness framework can help uncover where opportunities exist and which ones have a strong chance of successfully taking hold throughout the organization. Such activities also foster collaboration and encourage buy- in, both of which should happen as early in the process as possible.
Third, pilots and proofs of concept
are good ways to learn about emerging technologies, but they are not the end goal. Instead, organizations should plan, design and build for final operating capabilities. Too often, amazing solutions never make it to production because leaders didn’t plan for the broader goal.
Boosting cyber defenses and employee buy-in
Today, cyber defensive postures are reactive, and network defenders are losing. The average intrusion is detected 200 days after the fact. On top of that, the cyberthreat environment is more complex than ever, and traditional tools and rule-based defenses don’t scale with a limited cyber workforce. Adversarial tactics change rapidly, and models that train off-line are slow to update, rarely perform well, and often provide outdated and incorrect alerts.
To address those challenges, Booz Allen Hamilton has developed scalable solutions that augment established cyber defenses
THE GOVERNMENT’S PIVOT toward the adoption of artificial intelligence is about more than learning a new technology. To be successful, agencies must ensure that their AI efforts are consistent with and supportive of their broader business, technical and security strategies.
In addition, agencies need to understand and plan for the significant positive and negative impacts that AI can have on an organization. Rather than getting caught up in the hype around broad ethical concerns, for instance, agencies should spend time
distilling them down to the tactical level as risks that can be monitored, managed and mitigated appropriately. And they must look for ways to encourage employees to embrace the technology.
3 steps to AI adoption
First, it’s important for agencies to take stock of their goals relative to AI. Do
they want to create efficiencies, improve productivity or reduce a backlog? All AI projects need to connect to the ultimate business goal or mission objective, and it is equally important that agencies develop
davooda/Shutterstock/FCW Staff
S-20 SPONSORED CONTENT