Rootcode AI | We build great AI

Introduction to AI in Governments

Welcome to Part 2 of our blog series about the potential of E-governance. In our previous article, we delved into the various models of e-governance: G2C, G2B, G2E, and G2G. We examined how countries have implemented these models, facing both challenges and successes along the way. We highlighted the significant benefits of e-governance, such as increased efficiency, transparency, and accessibility of government services, alongside its potential to reduce corruption and operational costs.

In this article, we shift our focus to how AI can be utilized by governments to strengthen the way in which they provide services. Governments are increasingly recognizing AI's potential to elevate the quality and efficiency of their services. Through this article, we will explore a few examples of how AI can be strategically integrated across different e-governance models to streamline processes and significantly improve service delivery.

Enhancing G2C Services with Artificial Intelligence

The G2C (Government-to-Citizen) model plays a crucial role in e-governance by streamlining the interaction between government bodies and the public. Through the integration of AI, this model not only boosts the accessibility and efficiency of services but also enhances the user experience. AI can automate routine tasks, provide personalized service around the clock, and make government operations more accessible. Let's explore how AI is redefining G2C interactions through specific impactful use cases.

I. Citizen-Centric Public Policies Through AI

AI stands as a critical tool for governments aiming to shape more effective, citizen-centric public policies. By harnessing the power of AI, governments can engage in policy-making processes that are directly informed by the voices of their citizens.

In this modern digital era, citizen’s concerns are echoed vastly through their online presence. Moreover, government initiatives such as public consultations and survey responses contain data rich in multiple public opinions. Such publicly available data can be gathered to analyze large volumes of public sentiment data using language understanding algorithms. By evaluating the data from residents of specific areas, governments can identify concerns and needs before policies are formulated. This approach allows for a more informed and responsive policy-making process that considers the real-time moods and opinions of the public. Government officials can use AI tools to analyze feedback received through social media, detecting emotions such as fear, anger, happiness, and satisfaction. This analysis helps in informing adjustments in policies to better meet the needs and expectations of citizens.

A crowd-sourcing tool in Belgium developed by CitizenLab, a civic technology company, enabled decision-makers to analyze the citizen-contributed data in this platform effectively. In early 2019, amid protests against inadequate action on climate change, this platform collected over 1,700 ideas and received 32,000 votes from citizens. Leveraging advanced analytics, the platform helped distil these contributions into 15 priority policies. These policies were then presented back to the public for voting, ensuring that the final decisions reflected the community's priorities and insights.

II. Customer Support and Virtual Assistants in G2C Services

In today’s digital age, people expect rapid and efficient responses to their inquiries and requests. Traditional methods often fall short because of long waiting times and inconsistent service quality. This is where AI-driven technologies come in handy, fundamentally changing how citizens interact with government services.

AI-powered virtual assistants and chatbots are at the forefront of this transformation. They use Natural Language Processing and most recently, Large Language models to understand and respond to user queries with high accuracy. They are designed to provide 24/7 support, efficiently guiding users through complex bureaucratic processes, answering frequently asked questions, and assisting with form submissions. The implementation of Conversational AI Assistants enhances the efficiency of customer service operations, reducing the burden on human employees and ensuring that citizens can access necessary information and services at their convenience. More importantly, this presents an intuitive interface for citizens with relatively low digital literacy too.

For example, Emma is a virtual assistant that helps users navigate the USCIS (U.S. Citizenship and Immigration Services) website, providing information swiftly and efficiently. Trained using language understanding models with retraining pipelines, Emma is able to learn continuously and interact in both English and Spanish, making it accessible to a broader audience. Its capabilities include directing users to appropriate sections of the website, answering questions in real-time, and even speaking to users when the audio feature is activated. By handling an average of 1,550 queries, this AI Assistant significantly reduces the workload on USCIS staff and enhances the overall user experience by providing immediate, reliable assistance.

On the other side of the world, the AI assistant RAMMAS launched by the Dubai Electricity and Water Authority (DEWA) plays a critical role in facilitating interactions between citizens and various governmental services. RAMMAS assists with tasks such as bill payments, application tracking, and job applications, streamlining processes that traditionally required more time and physical presence. This AI tool is part of Dubai’s broader initiative to digitize and improve its governmental operations.

2. Enhancing G2B Services with Artificial Intelligence

The primary goal of the G2B (Government-to-Business) e-governance model is to promote better communication and cooperation between government agencies and the business sector. Businesses are highly valuable in contributing to a country’s economic development. As such, governments can benefit deeply from systems that promote communication and collaboration between government entities and business enterprises. This model includes initiatives such as online business registration, licensing and permitting systems, and procurement portals, which are designed to minimize administrative burdens, simplify the process of conducting business, and promote economic growth. By integrating Artificial Intelligence (AI) into these systems, the potential to transform and enhance these systems grows significantly.

I. Signaux Faibles: AI-Driven Initiative to Predict Business Distress

Historically, the delay in recognizing signs of distress within businesses has been a critical challenge. Early detection is important for effective intervention, which can significantly alter struggling businesses' outcomes. Their health is crucial for a country's economy, as they are the primary drivers of employment, innovation, and economic growth. When businesses fail, it can lead to job losses, decreased consumer confidence, and a reduction in economic activity, all of which can have ripple effects throughout the economy.

While government bodies seek to support companies in difficulty, the main challenge lies in identifying early signs of business distress. This is crucial in providing effective and quality interventions by the state for struggling companies. The integration of AI can facilitate the process of detecting such signs. The "Signaux Faibles" (which directly translates to “Weak Signals” ) project was launched in France as a collaborative effort between the French Ministry of Economy, Finance and Recovery, and other governmental bodies for this purpose. I.e The system proactively uses AI for the timely identification of financially at-risk businesses to facilitate state-targeted remedial actions to prevent business failures.

At its core, the system uses statistical and machine learning methods to analyze and process large information quantities such as past data, administrative data and the trajectories of businesses that have defaulted, to signal the risk of entry of a business into financial difficulties.

The administrative data includes economic, financial, and activity-related information. This diverse and complementary data set is crucial for detecting long-term trends, including debt and equity issues, and alerting to cash flow tensions or underactivity. The detection is done by a supervised learning model that processes vast amounts of data to provide statistical predictions of defaults up to 18 months in advance for companies. Once the predictive analysis is completed, a detection list of companies at risk is generated. This is then shared with various partner administrations, enabling the activation of their respective support mechanisms.

The "Signaux Faibles" initiative started as a small-scale project and has since scaled to a national level following the success of its early implementations. The national rollout was solidified through a partnership among several public institutions, ensuring a coordinated and comprehensive approach to supporting businesses in distress. The system has been continuously refined and expanded, incorporating additional features to enhance its effectiveness and subsequently support governments in fostering a more resilient business environment.

II. Digital Regulatory Reporting: Automated Compliance and Regulatory Reporting using AI

Regulatory compliance is a critical aspect of business operations, ensuring that companies adhere to laws, regulations, and guidelines relevant to their industry. Compliance requirements can range from financial reporting and environmental regulations to data protection and workplace safety. Effective regulatory compliance not only helps businesses avoid legal penalties and reputational damage but also promotes ethical standards and operational integrity.

Regulatory reporting, a crucial component of compliance, involves the systematic submission of data to regulatory bodies to demonstrate adherence to these standards. For regulatory compliance to be effective, a collaborative relationship between businesses and government is essential. Governments establish the regulatory framework and provide oversight, while businesses are responsible for implementing and adhering to these regulations. This cooperation ensures that regulations are practical and enforceable, supporting a fair and transparent business environment.

Traditional methods of regulatory reporting are often manual, time-consuming, and prone to errors. Businesses must compile vast amounts of data, which involves significant labor and resources. This means, as regulators, government authorities don’t always receive consistent or high-quality data. Integration of AI can support in addressing some of these challenges. The Financial Conduct Authority (FCA) in the UK along with other government bodies conducted initiatives on “Digital Regulatory Reporting” (DRR), a transformative solution to automate the process of regulatory reporting.

The DRR system automates the data collection task from various sources within financial institutions. Examples of data include transaction records, compliance reports, and other relevant information. The integration of data from various systems ensures that all necessary information is consolidated into a single, cohesive platform. The consolidated data is fed into a predictive model to forecast for potential compliance breaches based on the past behavior available in the records. Natural Language Processing (NLP) is also used to analyze textual compliance documents and communications to ensure compliance. The analyzed data is then used to generate regulatory reporting automatically. The pilot phase of this initiative reports a 30% improvement in data accuracy with Burges Salmon report estimating that businesses can save up to 20% on compliance costs through DRR. The success of the FCA's DRR initiative demonstrates the transformative potential of AI in regulatory compliance. Businesses have experienced significant benefits, including reduced compliance costs, improved accuracy, and real-time monitoring capabilities. The importance of G2B interactions cannot be overstated in this context. Effective G2B communication and collaboration are essential for creating a regulatory environment that is both rigorous and manageable. This integration not only supports businesses in adhering to regulations but also enables governments to maintain robust oversight, thereby promoting economic stability and growth.

3. Enhancing G2G Services with Artificial Intelligence

The Government-to-Government (G2G) e-governance model emphasizes the importance of fostering collaboration and synchronization across various government levels and departments and also government-to-government collaboration between different countries. By incorporating inter-agency data-sharing platforms and policy coordination mechanisms, the G2G model aims to streamline administrative processes and achieve greater uniformity and coherence in policy implementation. This approach not only improves the efficiency of government operations but also enhances the overall quality of public governance.

I. Smart Camera Automated Detection of Mobile Phone Use While Driving

The use of mobile devices while driving is a significant safety concern, contributing increasingly to road accidents due to driver distraction. For instance, In the United States alone, distracted driving claimed 3,142 lives in 2019, according to the National Highway Traffic Safety Administration (NHTSA). Traditional enforcement methods, which typically involve police officers stopping drivers manually, are effective but require extensive manpower and can be inefficient. Meanwhile, distracted driving habits are increasingly posing a risk to road safety.

To address this issue, the Dutch Public Prosecution Service (DPPS) has implemented an innovative solution involving smart cameras that automatically detects drivers using mobile devices. These cameras capture a photo of every passer-by. Photos are taken from a diagonal angle, capturing the driver’s actions without revealing their face, focusing instead on the presence of a mobile device and the vehicle's license plate. Such images respect privacy while ensuring accountability. The image is then fed into an image recognition algorithm to accurately detect the presence of mobile devices in the hands of drivers. The model is trained in these detection tasks under various lighting and weather conditions, ensuring reliable performance day and night. Therefore, If a driver is detected using a handheld device, the camera automatically forwards the evidence to the Central Judicial Collection Agency (CJIB). The DPPS is responsible for the prosecution and legal aspects, while the CJIB handles the administrative processing and fines highlighting how two different authorities are able to work seamlessly in collaboration through this technology.

This automated system has been proven to significantly enhance the enforcement capabilities of the Dutch Public Prosecution Service by reducing the manpower needed for on-the-spot checks and increasing the coverage as these camera equipment can be placed at a different location every day. During testing, a significant number of around 400 violations were detected with just two cameras, marking the effectiveness of integrating AI.

Screenshot 2024-06-11 at 12.28.15.png Photo caption: Cameras used to enforce handheld phone use behind the wheel. Source: (Click here)

ii. ATO's AI in Tax and Fraud Detection

The Australian Taxation Office (ATO) has embraced artificial intelligence (AI) to enhance its tax and fraud detection capabilities, marking a crucial step in its operational activities. To put this into perspective, the ATO's AI system has been instrumental in addressing the largest Goods and Services Tax (GST) fraud in Australia's history, successfully stopping $2.5billion in fraudulent payments and identifying more than $530 million of unpaid tax bills. These impressive results demonstrate the effectiveness yielded by AI’s implementation at an industrial scale.

The ATO utilizes AI to analyze large datasets and uncover insights that are impossible for humans to identify. This technology has become a powerful tool in the agency's efforts to recover nearly $45 billion in unpaid taxes owed by Australians. Additionally, ATO shares that it used deep learning models to identify $295 million in superannuation guarantee underpayments. Superannuation, a mandatory system in Australia, requires employers to contribute a portion of an employee's earnings into a retirement savings fund. Ensuring these contributions are made correctly is crucial for the financial security of Australians in their retirement. Interestingly, the natural language understanding models in this system searched through leaked documents, such as the Panama Papers, to detect $242 million owed by tax evaders since 2018. The system alerted its analysts to the most valuable documents and showed where and what to look for in the information to pinpoint tax cheats. Future enhancements to the system include the use of supervised learning with gradient-boosting algorithms to identify rapid evolution in GST fraud behavior. Given the sensitive nature of analyzing tax data for fraud detection, the ATO addresses ethical concerns by using AI as a supportive tool, ensuring that final decisions are made by humans.

The success of the ATO’s AI initiative is not solely due to the technology itself but also because of effective G2G collaboration. The ATO works closely with other government agencies, including the Australian Federal Police and the Australian Securities and Investments Commission (ASIC). This collaboration involves sharing data, intelligence, and resources to ensure a cohesive approach to tackling tax fraud. This clearly highlights that the significant advancement in governmental operations not only demonstrates the power of AI integration but also the power of G2G collaboration.

4. Enhancing G2E Services with Artificial Intelligence

The Government-to-Employee (G2E) e-governance model aims to enhance the operational efficiency within government agencies. This model has been pivotal in transforming employee self-service portals, online training programs, and performance management systems. These initiatives aim not only to increase the productivity of public sector workers but also to boost their job satisfaction and engagement.

i. Streamlining Restaurant Health Inspections with AI

Traditionally, the process of scheduling health inspections for restaurants has been manual and not always efficient, leading to potential oversight of establishments that may pose significant health risks. In many regions, the inefficiency of conventional inspection processes has led to substantial delays and oversight, adversely affecting public health. For instance, before the implementation of AI-driven systems, it could take several months to schedule and conduct an inspection, during which time potential health risks could go unchecked. This delay is problematic, particularly in the restaurant industry where food safety is paramount.

The General Food Directorate of the Ministry of Agriculture and Food of France was responsible for launching the "Food AI" (IAlim) in 2019 to assist in targeting restaurant health inspections more effectively. Based on data gathered from sources such comments and ratings posted by consumers on digital platforms such as TripAdvisor and Google, which includes over 10 million comments, and historical inspection results, a predictive algorithm is trained to generate a list of outlists identified as priority for risk of violation based on poor reviews or flagged issues.

AI's ability to process and understand human language allows it to go beyond mere numeric ratings and delve into the nuances of customer reviews. By using advanced sentiment analysis, AI can detect underlying emotions and concerns that may not be immediately apparent through traditional metrics. This comprehensive understanding enables "Food AI" to provide a more accurate assessment of a restaurant's potential health risks. For example, a series of reviews mentioning symptoms of food poisoning or poor hygiene practices can alert inspectors to serious issues that require immediate attention. The AI system can weigh the severity and frequency of such mentions, offering a prioritized list of establishments that need urgent inspection.

"Food AI" (IAlim) demonstrates how AI can enhance the G2E e-governance model by improving operational efficiency, ensuring timely and effective inspections, and safeguarding public health. As people continue to share their experiences across various platforms, AI's capability to aggregate and analyze this information becomes increasingly valuable. "Food AI" exemplifies how technology can harness the collective voice of consumers to improve safety and quality standards in the restaurant industry.

ii. UK’s Civil Service Learning Platform

The United Kingdom’s Civil Service Learning (CSL) platform exemplifies a forward-thinking approach to Government-to-Employee (G2E) interaction. The platform is a centralized hub offering a wide range of courses and resources designed to meet the diverse training needs of UK government employees. By providing government employees with comprehensive online training and development opportunities, the CSL platform enhances workforce skills, promotes continuous learning, and fosters a culture of excellence within the public sector.

The CSL platform is built on robust e-learning technologies that ensure a seamless and interactive learning experience for users. As such, the platform utilizes recommendation algorithms to personalize learning paths based on individual user profiles, job roles and career aspirations, thus creating a relevant and targeted training. This personalization has become the norm in today’s era of information consumption. Therefore, such integrations can enhance the learning engagement and outcome of the users.

By leveraging modern e-learning technologies and fostering a culture of continuous improvement, the CSL platform not only enhances individual competencies but also contributes to the overall efficiency and effectiveness of the government. The success of this G2E initiative demonstrates the transformative potential of digital learning solutions in the public sector.

Conclusion

The adoption of AI in governance strategy has the potential to positively impact how governments deliver services making it smarter and more data-driven. From improving citizen interactions through AI-driven public policies and virtual assistants, enhancing business environments with predictive tools and automated regulatory reporting to facilitating inter-agency collaboration and effectively improving operational efficiency, the versatility in the integration of AI is not only undeniable but also transformative. The future of AI in e-governance holds promise for more agile, data-driven decision-making processes that can elevate the quality of public services and foster a more inclusive, efficient, and responsive governance system. As we continue to witness these technological advancements, the potential for AI to transform government services grows ever more promising.

At Rootcode AI, our mission is to help governments and enterprises craft end-to-end AI strategies. Recently, we began collaborating with the Estonian Government to build an AI model training and deployment platform for key government services. This partnership is a testament to our commitment to driving innovation and efficiency in public administration through cutting-edge AI solutions.

Talk to us to collaborate on your next project together!

The Role of Artificial Intelligence in Modernizing Government Operations and Services

Table of contents: