AI in Public Sector: 3 Recommendations for Governments

As the world becomes increasingly reliant on artificial intelligence (AI), governments are scrambling to keep up. The promise of AI is alluring: improved efficiency, enhanced decision-making, and better services for citizens. But, as governments attempt to implement AI technology in the public sector, they're facing a multitude of challenges that threaten to wreck their efforts. 

In May 2024 Government Tomorrow Forum (GTF) surveyed 120 AI experts from academia, business, and consultancy to find out the most pressing challenges that are faced by the governments for AI implementation. The number one drawback appeared to be social, ethical, legal, and societal concerns (SELAS concerns). Participants of the survey agreed  that it directly influences society’s trust in AI technology, and, as a result, the overall volume of AI implementation.

This article outlines recommendations for political leaders to help them address the uprising difficulties.

Develop Clear AI Guidelines

The concerns are the foundation upon which AI is built. Unmitigated bias in computer algorithms can lead to discrimination, privacy breaches can erode trust and lead to mass uncertainty, slowing down technological development. It's like building a house on shaky ground — it's only a matter of time before it comes crashing down.

No one would enjoy living in a world where AI-powered job recruiters discriminate against certain groups, or where AI-powered healthcare systems misdiagnose patients due to biased algorithms. The consequences are dire, and governments must take SELAS concerns seriously to ensure AI benefits everyone, not just the privileged few. Some of the primary recommendations are:

• Promote public engagement and dialogue

By fostering open discussions with citizens, experts, and industry leaders, governments can create a culture of transparency and trust, ensuring that concerns are heard and addressed proactively.

• Invest in AI ethics research

Governments can support research initiatives focused on understanding and mitigating societal risks associated with AI. This includes research on bias detection and removal in algorithms, as well as the potential societal impacts of job automation. 

Create Data Frameworks

Governments often grapple with those who own the data collected by AI systems, especially when it involves citizen information. This requires establishing clear data ownership frameworks. It will help to define who owns data collected by government agencies and how it can be ethically used for AI development. It is crucial because high-quality data is the fuel for AI. Without clear ownership and governance, there are risk issues like data privacy violations or limited access to valuable statistics. Additionally, governments can develop protocols for secure data sharing between public and private entities for legitimate AI development purposes. This fosters collaboration while safeguarding privacy and security. 

Educate Citizens

Public trust is essential for the successful adoption of artificial intelligence. If citizens fear job displacement, algorithmic bias, or manipulation, they may resist AI implementation. To address these concerns and ensure public buy-in, governments can take three key steps: 

• Launch national AI literacy campaigns 

Governments can develop educational programs to inform citizens about AI basics, its potential benefits and risks, and its applications in various sectors. This empowers citizens to participate meaningfully in discussions about data policy and technological development.

• Invest in AI hard skills development 

By providing training programs to equip people with the skills needed to work with and adapt to AI technologies. This helps address concerns about job displacement and prepares the workforce for the future.

• Promote transparency in AI decision-making

Governments should ensure transparency in how AI systems are used in government services. This can involve providing clear explanations on how algorithms make decisions and offering appeal mechanisms for unfair outcomes. 

Success Stories 

Despite the numerous challenges, there are several successful cases of AI implementation in the public sector. This proves that progress is achievable, and we can look to pioneering examples for guidance.

Estonia, for instance, pioneered the e-Government System, utilizing AI-powered chatbots that answer citizens` questions and provide personalized services. Their X-Road platform allows secure data exchange between government agencies and businesses. They leveraged blockchain technology for identity verification and educating citizens on how AI is used. In addition, their comprehensive awareness campaigns dispelled misconceptions, fostering public trust and expanding global digital business opportunities.

Another notable example is the New York City Department of Social Services (DSS). The DSS, having a limited amount of resources, faced the challenge of processing a high volume of applications for the Supplemental Nutrition Assistance Program (SNAP). To address this, they implemented an AI chatbot online self-service designed to answer basic questions, guide users, and reduce the need for in-person visits. By combining advanced AI technology with human assistance, the New York City DSS successfully improved service delivery, reduced processing times, and built greater trust within the community. The system underscores that it focuses on clear explanations about data collection and use. 

Where to start

To get meaningful results and avoid getting lost in routine tasks, it is important to clearly define problems with measurable outcomes in the beginning. The task should address a specific, well-documented issue, enabling the project to demonstrate its impact and effectiveness after a while. For instance, automating repetitive tasks or implementing fraud detection systems are clear, measurable objectives. This clarity makes it easier to evaluate success and gain approval for broader adoption.

By working together and focusing on clear goals, national and regional governments can unlock the full potential of AI, addressing the challenges and maximizing the impact from this transformative technology.