AI levels the playing field in the public sector
This post has been co-authored by Thanos Tatsios (Advisory Data Scientist, IBM) and Steven Hintze (Chief Data and Product Officer, Arizona Department of Child Safety).
One way a company conveys its ideas for behavioral guidance is through policy. Deborah Stone is a renowned scholar who knows a thing or two about policymaking. She has taught at multiple universities, including Brandeis and MIT, and wrote a book titled, “Policy Paradox.” She has called the struggle over ideas the essence of policymaking:
Ideas are a medium of exchange and a mode of influence even more powerful than money, votes, and guns. Shared meanings motivate people to action and meld individual striving into collective action. Ideas are at the center of all political conflict. Policymaking, in turn, is a constant struggle over the criteria for classification, the boundaries of categories, and the definition of ideals that guide how people behave.
When policymakers get to work there inevitably is intentional and thoughtful consideration into the implications of the words being written in a way that is sufficiently broad for guidance, yet effectively honed for the intended impact, they aren’t the only ones who need to understand them. All the subject matter experts will provide their contributions, the legal team will ensure alignment with relevant laws or rules, the operational heads will contemplate the impacts to the flow of work, etc. and in the end a product is created and posted for all others to consume and follow as the way forward.
These bright minds will then deal with the inevitable next human behavior — questions rolling in. That’s because when a company writes a policy document, the intended audience can vary depending on the type of policy and its purpose. The stakeholders are potentially a wide range population for a company’s policy documents and may include:
1. Employees
2. Management/supervisors
3. Shareholders/investors
4. Citizens/customers/clients
5. Regulatory bodies/legal authorities
6. Business partners/vendors
In itself this presents a challenge on ensuring accessibility to the necessary information by impacted parties. Not only in terms of accommodations for individuals who need to access the information without the ability use eyesight, but also accessibility in terms of making the subject matter understandable to the impacted party. Not to mention the difference in perspective and need from each group as an entity.
Understanding a company’s policies can be tough for anyone. This includes critical and non-critical policies, whether created by the company or required by law. Companies have diverse teams. These teams include people of different ages, cultures, education levels, and experiences. This diversity makes it even harder to grasp complex policies. Policies are meant to guide. However, they can also hinder clear communication. This creates gaps in understanding across the team. Policies often cater to those with specialized training or legal knowledge. As a result, most people might miss out on important details and implications.
The inability to communicate effectively can perpetuate inequality across the workforce, limiting the number of well-informed and well-connected people capable of carrying out an organization’s mission. With the advent of artificial intelligence (AI), we can seek to level the playing field and enable further understanding of policy information for everyone within a company. We can achieve democratized access to vital behavioral knowledge.
The question now is: How willing are organizations to boldly challenge the status quo and unleash the power of AI to make the policy understanding more human-centered and give everyone a fighting chance to thrive in a complex world? And how can you address the risks blazed across the media of hallucinations and glue on pizza? Albeit the later question may have the same risk of a well written policy that is misunderstood by it’s intended audience that leads to an unintended outcome.
How AI can act as an equalizer
The rapid advancement of AI technology has started to change the workforce dynamic to serve as an equalizer. Companies can empower each persona type with an ability to better understand the policies to better guide behavior and outcomes.
A sample of areas where AI can transform policy management:
- Accessibility. The increasing use of natural language processing (NLP) is transforming the way we interact with complex legal language, making it plain and accessible. By leveraging NLP, AI assistants can swiftly analyze extensive policy documents, distill the key points, and present the information in a clear, easy-to-understand format. When people are presented with information that can be consumed and digested, they are more likely to quickly grasp the essentials details without being overwhelmed by convoluted terminology. By making policy information more accessible, AI fosters greater transparency and accountability in government, reducing the risk of policy manipulation and abuse.
- Policy navigation. By boosting efficiency and document navigation, AI eliminates the need for time-consuming manual searches and interpretation of complex texts. Tuned AI assistants can swiftly locate policies and provide source information to each persona type. Going beyond simple search, comprehensive knowledge graphs can unveil the intricate connections between policies, visually displaying how an individual policy complements or interacts with another policy. To make this complex information digestible, we can employ intuitive visualization tools such as interactive dashboards and clear graphical representations. Bu turning policy navigation into something that is visual, the underlying narratives can be turned into something that is intuitive and engaging, allowing users to quickly grasp the big picture as well as the details.
- Rapid implementation. Policy professionals can now harness AI tools to quickly adapt to new and updated policies seamlessly translating them into operational changes. This accelerates and enhances policy implementation and compliance among government agents (or commercial organizations, too) significantly reducing the risk of being ill-informed, violations and penalties. By delivering accessible policy knowledge and streamlining processes, AI empowers confident decision-making and frees up valuable time for core tasks, leading to substantial cost savings. Moreover, AI is now adaptable and capable of managing the ever-growing volume of policy documents with ease.
Arizona State, Department of Child Safety
The Arizona Department of Child Safety (ADCS) is dedicated to protecting Arizona’s most vulnerable populations, including children and families. To achieve this mission, ADCS relies on effective policy management, ensuring that policies are up-to-date, accessible, and easily understood by all stakeholders. ADCS has turned AI into a useful, productive, and boring approach to policy management. The approach is not to chase the hype-cycle, but to be reasonable, responsible, and effective. We have recently introduced an AI assistant, designed to navigate through policies with ease, providing instant access to relevant information and streamlining policy management. Previously, this would require new employees and existing employees to know exactly what complex child welfare terms to enter into a search bar, and to select the correct search result, and to know that they may need to continue searching for related policies. It’s not impossible to do, but it is not user friendly.
We sought out to both create safe and secure infrastructure using a trusted high quality data set (our policy manual) as a means to increase the data quality of information entered into the system by providing just-in-time access to necessary policies and procedures while working within our CCWIS systems. As a side benefit it reduces time it takes for ADCS employees (and in the future the public) to interact surface relevant policies, ensuring that everyone has the information they need to make informed decisions.
The CCWIS AI assistant, will be responsibly phased in starting at a beta test with a small set of users, and then to be scaled up to provide equal access to policy information for all stakeholders, including employees, partners, and the public, regardless of their technical expertise. The AI assistant provides equal access to policy information for all stakeholders, including employees, partners, and the public, regardless of their technical expertise. The AI assistant promotes transparency by providing easy access to policy information, reducing the risk of errors or misinterpretation.
As ADCS continues to leverage the power of AI, we can expect to see even more innovative applications in policy management and beyond. The AI assistant is just the beginning, paving the way for a more efficient, transparent, and effective child welfare service.
Conclusion
The advent of AI holds the potential to reform our interaction with complex policy information, democratizing access to this vital knowledge and leveling the playing field for all. By bridging the knowledge gap, AI can empower individuals from diverse backgrounds to effectively utilize policies, promoting greater transparency and accountability while driving equity and fairness. If we make the choice to use AI, we can transform how citizens or other personas engage with policy, enabling them to understand intricate issues, make informed decisions, and advocate for their interests. By leveraging AI’s power, we can create a more informed and engaged community, fostering a more democratic and equitable society. In an era of equality, equity, and justice AI has the potential to become an effective communication tool.
