1) AUTOMATES THE RECRUITMENT PROCESS

• AI could benefit the government human resource department by streamlining candidate sourcing and automate interview scheduling.
• It could also promote diversity and help the government avoid negative criticism during hiring.
• According to Govloop, relying on manual interviews and assessments to hire candidates, is guesswork because interviews do not cover everything the organization will want to know. A.I. will take the guesswork out of the way and allow governments to hire the right people by using algorithms to analyze job and performance-related data.

2) RELIABLE TAXATION

• The IRS can use AI for financial analysis on multiple transaction mediums for the purpose of taxation.
• The use of AI for taxation will benefit the government by providing efficiency and fairness and this will curb tax evasion and double taxation.

3) EASE TRANSPORTATION

• The transport department use of AI in the transportation system will reduce traffic congestion and improve the punctuality of its public.
• According to Deloitte Insights, AI can cut travel times by 25% and idling times by 40%.
• According to the secretary of transportation, the use of AI such as autonomous vehicles will decrease traffic fatalities and improve mobility.

4) ADVANCE SECURITY AND SAFETY

• Artificial intelligence can help advance security by predicting natural disasters and aiding in rescue efforts.
• Artificial Intelligence can also relay warnings of external threats to the government.
• The DoD is focusing on the use of AI in challenging cases that can benefit from AI such as perception, maintenance, humanitarian assistance, and disaster relief.

5) HELP IN LAW ENFORCEMENT

• The police department can benefit from AI by using it to clear cases that can be automated by bots.
• The city of Chicago is trying to prevent violent crimes before they happen using AI. The city's predictive analytics unit uses 911 call data to run spatial algorithms to identify where violent crimes are likely to happen.

6)DETECTING IRREGULAR FINANCIAL ACTIVITIES

• Government's use of AI will benefit it by detecting irregular financial activities such as money laundering and fraud.
• AI can also be used to detect terrorist financing. This will promote peace and greatly benefit the government.

7) IMPROVE INFRASTRUCTURE

• AI can be used to help optimize infrastructure, built smart cities, and generate and save energy.
• The use of AI will help the government to improve decision-making by providing accurate information needed to design, construct and operate infrastructure assets.

RESEARCH STRATEGY

#1: JOB ELIMINATION

• One major detriment to using AI in government is the likely subsequent loss of jobs. AI use has the potential to progressively destroy first blue-collar jobs, then white-collar jobs, and finally professional jobs.
• Unemployment is a major political issue, and therefore could politically threaten the administration in power.
• Finally, while AI has the power to eliminate many jobs, it only creates jobs for highly skilled, highly paid individuals, which could increase inequality. As another major political issue, increased inequality could threaten the administration in power.

#2: LACK OF TRANSPARENCY

• The code that makes up AI is far too complicated for most people to understand. This complexity is a detriment to the government because it threatens its ability to be transparent to the public, which is an ethical and legal issue.
•  "Black-box" AI are AI technologies so complex that even the engineers who created them struggle to explain them. In this report, several AI experts assert that government agencies responsible for criminal justice, healthcare, welfare, and education should stop using black-box AI because of their lack of transparency.

#3: POOR DATA AVAILABILITY

• For AI to work, it needs large quantities of rich data. Data that is not clean and complex can be very risky. For a couple of reasons, many branches of the US government do not have the data necessary for functioning AI.
• Because government agencies compete for funding, they are also often reluctant to share data openly. This limitation means that government data is not very rich, because of the lack of integration.
• Because of privacy concerns, some data may not legally be allowed to be collected and used by AI.
• Many agencies still do not prioritize data collection and analysis as a way to make decisions, and thus lack usable data.

#4: DIFFICULT IMPLEMENTATION AND INTEGRATION

• Much of the IT infrastructure in the government is outdated, which means that updating the systems so that they could incorporate AI would be costly and difficult.
• To effectively use AI systems, the government employees working with the systems would have to receive a high level of re-training, which could be costly and time-consuming.

#5: EXPENSE OF AI PURCHASE

• Because the IT infrastructure in government is outdated, it is expensive to maintain, and therefore allows a limited budget for AI implementation.
• The US government has a long history of IT projects that run hundreds of millions of dollars over budget, and the root causes of this phenomenon are unlikely to fade away soon. So, an expensive AI implementation would be at risk.
• Partnering with the private sector in order to introduce AI to government procedures has been excessively expensive in the past.

#6: RISK OF PRIVACY VIOLATION AND DISCRIMINATION

• Using AI systems could threaten the US government because of the legal risk that the systems could violate privacy and propagate discrimination based on a protected category.
• An algorithm built to predict the likelihood that a convicted criminal would commit another crime demonstrated a bias against black people, despite actual results that contradicted the pattern of predictions. This article further expands on the Fourth Amendment implications of using big data and predictive algorithms in policing.

#7: ACCOUNTABILITY AND SUPERVISION ISSUES

• Similarly to the lack of transparency, AI in government poses a risk in that it creates accountability and bias problems.
• AI technologies that operate without human supervision cannot make "judgment calls" the way that humans can.
• For this reason among others, AI experts signed this letter asking to ban the AI tool of "autonomous weapons," which can identify and attack targets without human intervention.

#8: SECURITY RISK

• Integrated AI systems that connect different components of government pose a security threat because hackers have easier access to diverse information from hacking into one system.
• Researchers from NYU built a hack into the AI systems of the multi-billion dollar cloud service market, which includes Amazon, Microsoft, and Google, which allows them to trick the AI. If this kind of hack occurred in government AI, the security implications could be grave.

RESEARCH STRATEGY