Cloud-based AI for government
As the founder of the Dcode42, accelerator for companies addressing current or potential future government need, Meagan Metzger talked about how her company's partnership with Amazon Web Services will help speed the adoption of artificial intelligence and machine learning for problem solving in government.
Government Cloud Insider spoke with Metzger about the role of AI in government and ways cloud-based AI can help government solve challenges. The interview has been edited for length and clarity.
GCI: What government challenges do you see AI solving?
Metzger: There are many areas that AI can impact in the government space. They range from things as simple as automating human-intensive processes, such as scheduling meetings and payroll. And then on the opposite end of the spectrum, AI can play a pretty huge role in cyber defense and health, such as predicting patient survival rates. It really spans a wide spectrum across the government and could play a pretty huge role in increasing efficiency.
GCI: How does AI help with the decision-making process?
Metzger: It can correlate information and see patterns across large datasets that it might take humans quite a while to process, and a lot of times it can do that with more accuracy. AI and machine learning can look at information that's presented and not just understand that A and B are related, but how they are related and the intricacies between the different data points.
GCI: What are the advantages of cloud-based AI?
Metzger: Amazon in particular has a very strong interest in AI, not just for applications that run on their platform, but they're also constantly evolving new toolsets for software providers and developers to increase machine learning and advance the different capabilities that their solutions can provide.
When we teamed up with Amazon for our accelerator program, they wanted to focus on AI because they see that there's such a big impact that AI can make across several use cases. They believe that AWS can play a huge role in enabling private-sector AI solutions to help solve government challenges.
I think these toolsets and the cloud infrastructure are going to be critical as AI capabilities enable a lot of these solutions.
GCI: What are some of government challenges AI can help with?
Metzger: I think detecting fraudulent activity is a primary use case across the government. AI can correlate lots of different information -- from the accounts being accessed, to who is accessing them, and the users’ relationship to watch lists or other red flag indicators. It can take what's normally siloed information and connect it together.
GCI: What about for cybersecurity?
If you think of cyber holistically, there are some interesting applications around not just the technical aspects of cyber attacks -- being able to identify when the system has been targeted -- but also looking at sentiment analysis. Looking at changes in the subject matter and participants in communications -- and processing the more quantitative tasks required for that kind of analysis -- can better detect a cyber threat.
GCI: How can AI support the military and the intelligence community specifically?
AI and machine learning could have a huge role in predicting where the next attacks are going to be. Some of that can be from just detecting different patterns in communications across some known threat vectors. Being able to put that prediction piece in place is huge and could really prevent a lot of bad activity from going on.
AI can also have much higher accuracy around things like image recognition. We get a lot of fantastic footage coming back from drones or on the ground from troops. Being able to identify and recognize what is going in those images could be huge indicators for different scenarios.
Leila Meyer is a freelance technology writer based in British Columbia.