How I used Rapid Prototyping to build 5 AI Applications in 10 days

On February 11th 2019, the White House issued an Executive Order launching the "American AI Initiative", this initiative was followed by the release of www.AI.gov on March 20th 2019. The American AI Initiative encourages Federal Agencies to explore and grow AI capabilities. The United States isn’t the only country exploring AI Technology. According to China’s State Council, by 2030 China’s AI industry could be worth $150 Billion.(Larson, 2018)
Developing AI technology
Developing AI technology can be a challenging endeavor. An effective AI solution is a combination of workforce, internal governance, and technology innovation. Building an AI workforce can be very expensive because there still aren’t enough Data Scientists, Data Engineers, and Analysts to fill positions. The reorganization of a goverment agency’s legacy data architecture to efficiently support Machine Learning Model development can take a long time and require a huge upfront investment. It’s possible that leaders may have to shift their data from on-premises servers to a cloud or hybrid based solution that can be efficiently served by Hadoop Clusters. There is also a challenge with bringing the current workforce up-to-speed so that they can make the most of the new AI technology. There will be a significant workforce training investment to get executive leadership, managers, and product developers to a place where they know when to initiate the development and deployment of Machine Learning Models to maximize business drivers, save operational costs, and develop new products. Developing a robust AI technology organization is more than just buying a platform and just letting it work. It will be a shift in investments, internal business processes, and organizational strategy.
Agile AI solutions
To make any significant change takes time. The process can be made smoother by achieving early wins. There are many inexpensive or free AI tools like bot builders, pre-trained Machine Learning Models, and AutoML solutions that can be used to quickly and inexpensively build AI app prototypes and production level solutions. Federal Agency IT solution implementation is notorious for being overly expensive and time intensive. In today’s world, time is a luxury that can not be taken for granted. Government Innovators and Technologists must develop new methodologies and processes to operationalize Artificial Intelligence and Machine Learning Technology solutions quickly and efficiently. This is the only way Government Agencies can remain competitive in the AI realm and properly serve the United States citizen.
My Personal Government AI App Challenge
As a proud veteran, I loved serving my country. Although I am far from my days in the Armed Forces, I still have a strong need to serve my country. I feel that I can still serve my country by utilizing my technical knowledge and skills. I am challenging myself to create 5 Government AI app prototypes in 10 days using Google Cloud Dialogflow and AutoML. I am doing this to show that it doesn’t take a long time or an expensive acquisition process to create prototype applications. These apps will be based on publicly available information. I will demonstrate how government agencies can use AI to help American Citizens more easily navigate their policies, FAQs, and other digital knowledge.
Artificial Intelligence App 1:
Smart Vet App
Description:
The app will help Veterans with quickly getting information on VA hospitals, programs, and filing disability claims. I built this app with Dialogflow’s multimodal functions. The app would be available on Smart displays and mobile.
Artificial Intelligence App 2:
Car Seat Checker
Description:
This app will help parents by ensuring that they have placed their child in the car seat properly. The parent will be able to take a picture of the child and text it to an app via Twilio and the app will tell the parent if the child is placed in the car seat properly. I built and trained a Machine Learning Model with Google Cloud AutoML on incorrect and correct images of children in car seats.
Artificial Intelligence App 3:
VA Hospital App Navigator
Description:
This app will help veterans with navigating VA hospitals in the pre-appointment phase, appointment phase, and post-appointment phase. I built this app with Dialogflow’s Twillio integration and it will be text based only with rich content.
Artificial Intelligence App 4:
GSA FAQ App
Description:
This app will work on Google Assistant Devices and will help citizens, partners, and vendors navigate the over 87,000 pages maintained by the GSA. I built this prototype with Dialogflow.
Artificial Intelligence App 5:
Veteran’s Opportunity App
Description:
This app will help newly transitioned Veterans find opportunity for post military career employment. Any vet would be able to call the AI on a phone number to ensure that even non-tech vets could get access to this resource. I built this prototype app was with Dialogflow’s telephony integration.
Concluding Remarks
Developing and deploying AI solutions in government agency can be technically challenging and expensive. However, buy-in, support, and product direction can be achieved by rapidly prototyping AI products and services. I highly encourage all Government Innovators to explore these techniques to bring AI to their agencies.
A write-up on each of the 5 AI Government Apps are at the following links:
Sources:
(1) Larson, Christina (2018, February 08). China’s massive investment in artificial intelligence has an insidious downside. Retrieved January 10, 2021, from https://www.sciencemag.org/news/2018/02/china-s-massive-investment-artificial-intelligence-has-insidious-downside