AI in public sector
Governments and public services organizations are undergoing a transition.
Today, the focus is on prioritizing citizens’ participation in decision-making. Governments are using real-time data to improve service delivery and citizen experience, optimize expenditures, and increase revenue sources. Technology can play a vital role in enhancing citizen outreach, delivering citizen services seamlessly, and addressing the priorities of governments and public services organizations.
Among the technological advancements, AI, as a field, has been researched and discussed for over 70 years. However, despite its potential, AI has been subjected to fluctuating interest, with communities talking metaphorically about the AI winter when people’s interest wanes due to unmet expectations.
Today, both public and private sectors are equally enthusiastic in adopting the technology. According to the initial AI mapping by the Organization for Economic Cooperation and Development (OECD), 50 countries (including those in the European Union) have introduced national AI strategies, with 36 having specific plans for the public sector. Most of these AI strategies follow similar themes, covering economic development, building citizen trust, ensuring ethical conduct, strengthening information security, and enhancing the talent pipeline. While there is no uniform legislation governing the use of AI in the public sector, many governments have launched national projects to utilize this technology.
Generative AI, a step ahead of traditional AI, is a powerful tool that can revolutionize the way governments operate. When used responsibly, it can be a force for positive change.
Areas of impact
The public sector stands to gain exceptional benefits by integrating AI into every aspect of their work and decision-making.
We believe AI will have immense potential to recalibrate and elevate citizen experiences (see Figure 1).
Hyper-personalized citizen services: Generative AI along with other digital technologies will enable government agencies to automate research and data analysis around citizens’ needs and preferences. Government agencies can, therefore, offer personalized, contextualized, and responsive services for a seamless citizen experience using demographic, infrastructural, social, and historical data.
Governments have long sought to overcome regional, ethnic, and social disparities in service delivery; generative AI offers a solution by delivering personalized services to citizens.
Increased engagement with citizens: Generative AI-powered chatbots and tools can optimize the quality of interaction and communication with citizens by providing contextual information based on their needs, answering queries (better than standard replies provided by a chatbot) by leveraging large language models, analyzing, and scanning government policies and documents in real-time, and collecting feedback. This can reduce employee workload and enhance citizen engagement.
Improved staff productivity and experience: Generative AI can serve as an able back-office assistant to government agencies. It can automate repetitive and time-consuming tasks involving processing, scanning, and validation against tedious policy documents using large language models, freeing up bandwidth to prioritize complex and strategic work. This can help reduce workload and increase productivity, leading to greater job satisfaction and improved outcomes.
Informed decisions backed by data and insights: Generative AI can improve public services by analyzing bulky policy documents, previous interaction history for similar scenarios, and data to identify trends and risks for better decision-making and proactive measures. This is especially crucial in public safety for quick threat identification and response.
As generative AI advances, we can expect impactful applications for enhancing governance and delivering top-notch services in various areas, including:
1) Healthcare: Generative AI can streamline processes and enable optimal clinician time utilization.
Scheduling assistant: Using generative AI to schedule appointments for general practitioners and specialists based on citizen’s preferences can reduce cancellations and no-shows, resulting in a more efficient and effective scheduling process.
Making clinical decisions: Generative AI can rapidly analyze medical records, diagnostic reports, and medical imaging to provide accurate diagnoses and treatment plans. It can also help promote telemedicine in areas with limited medical professionals.
Assessing potential pandemic risk: With rapid mutations in viruses, any new pathogen immune to existing medication can escalate into a pandemic-level crisis. Scientists and doctors can leverage AI models to analyze potential catastrophic developments, study their impact, and develop preventive measures.
Monitoring progress and compliance of key initiatives: Generative AI can assist in spreading helpful information and tracking progress and compliance of critical initiatives across regions by harmonizing and analyzing data and inputs received from multiple channels and departments involved (such as supply chain, enforcement, and monitoring teams). It can also send timely reminders to citizens for enrolling in health plans as part of open enrollments. Additionally, it can remotely monitor the aging population and provide assistive personalized care.
Researching: Generative AI is being used for drug discovery, genome profile analysis and development, monitoring and reporting adverse events, scenario simulation, and synthetic data preparation. Multiple initiatives are being prototyped around generative adversarial network model utilization for disease gene prediction with sequencing data and adverse event reporting.
2) Services delivery: Generative AI can improve the overall citizen experience.
Automating services: The technology can automate citizen services like appointment management and issuing birth, marriage, and vehicle registration certificates. Public services organizations have also started exploring interactive generative AI to automate responses to citizens’ needs for cases involving processing, scanning, analyzing, and evaluating bulky policy documents. An engaging interface combines text, voice, image, and video capabilities based on demographics and needs, improving satisfaction and reducing customer service workload.
Writing aid: Generative AI can also assist in creating abstracts, outlines, speeches, simple correspondence, memos, frequently asked questions, citizen guides, and legal documents such as contracts, briefs, and motions. While official communication would always require human intervention to verify accuracy, apply human voice, and ensure that the information is complete and accurate, generative AI, as a creative writing aid, can accelerate the process dramatically. It can light up the creative spark while reducing time-to-completion for common writing tasks.
3) Law enforcement and judiciary: The availability of real-time democratized data is critical for law enforcement agencies and judiciary.
Monitoring and predicting crimes: Such data is beneficial for monitoring and managing emerging law and order situations across jurisdictions, tracking the activities of perceived threats and high-risk personalities, and alerting law enforcement. Government agencies can also deploy AI-enabled cybersecurity solutions to scan, detect, and mitigate cyber threats.
Improving road safety: Generative AI can optimize traffic flow by examining real-time traffic data and predicting congestion before it happens. This can help reduce traffic jams and emissions and improve road safety. Generative AI can also analyze driver behavior using telematics, dash cams, and sensor data to detect accident sites, identify root causes of accidents, and provide personalized interventions and feedback to drivers involved and other commuters in general. Another potential application area is dynamic fines management and addressal of standard low-severity traffic violation tickets through an automated faceless interface.
Reducing case pendency: Generative AI can help reduce case pendency and manage complex litigation. The technology can summarize case law history, analyze evidence, assist courts, automate file scrutiny, and dispose of low-complexity cases to allow judicial officers and support staff to focus on core judicial aspects, enabling speedy justice. Generative AI can be leveraged effectively in arbitration and mediation scenarios to resolve corporate disputes by reducing the number of adjournments and sessions through the ready availability of all the relevant details. It can help build simulations to evaluate the probable impact of any award or negotiations on the parties in real time. Generative AI can help analyze bulky arguments of arguing counsels vis-à-vis different jurisprudence and interpretation styles for the bench or jury.
4) Budgeting, planning, and risk management: Governments diligently plan financial allocations during budgets.
Optimizing budgets: Generative AI can help governments optimize budgets by detecting unnecessary spending and allocating resources to meet citizen needs. It can also identify and evaluate risks and revenue loss, creating mitigation plans for public services.
Urban infrastructure planning and development: Urban infrastructure planning and development is tedious. It involves managing traffic, sanitation, waste, utilities, civic amenities, taxes, and constructing new residential areas and townships. Generative AI can assist urban planning through insights obtained from advanced simulations, enabling informed decision-making. Currently, this activity predominantly relies on multiple iterations of manual surveys through devices prone to human errors.
Governments must focus on four foundational pillars to capitalize on the benefits of generative AI.
A methodical approach toward adopting generative AI is key to reaping the benefits through accelerated cycles backed by strong executive sponsorship. The pillars (see Figure 2) include establishing a robust digital core, creating a data edge information fabric, strengthening information security, and promoting competencies.
Figure 2: Four foundational pillars
Establish a robust digital core: Governments and public services organizations should adopt digital technologies to modernize their legacy infrastructure. This will create a robust digital core for collaboration and enable generative AI use cases.
Create an information fabric for realizing data edge: Governments and public services organizations face challenges in managing their vast data estate for insights-driven decision-making. Establishing a data office can institutionalize information management standards and best practices across the entire data estate, enabling data harmonization for generative AI use cases.
Strengthen information security infrastructure: Governments have a fiduciary relationship with their citizens and must follow a zero-trust policy when handling citizens’ data. A robust information security framework is integral for realizing generative AI use cases.
Build strong governance and democratize competencies: Strong executive governance and sponsorship are essential for the success of potential generative AI use cases. Democratization of technological capabilities across departments can reduce rework and encourage optimal utilization.
Ethical concerns and the future
Governments and public services organizations sit on the largest pool of citizens’ data.
As data custodians, these agencies must be vigilant about the risks of data breaches, citizens’ privacy, and potential malicious use of technology. Using synthetic and dummy data or data available on public forums for model training can lead to data biases, damaging citizens’ confidence and having adverse political and legal consequences.
Considering this, governments and public services organizations may prefer their proprietary infrastructure instead of using a public or community edition of generative AI platforms. Models should be trained on real-world data in a controlled environment. A comprehensive set of AI regulations backed by a proactive, collaborative AI governance alliance both at the global and the national level is the key to ensuring constructive and responsible use of generative AI.