Generated AI navigation is at risk of public service delivery

Government agencies at the federal, state and local levels are obligated to provide cost-effective, reliable and equitable public services to all possible clients. Artificial intelligence tools may help agencies meet this responsibility by increasing the efficiency and ease of public service delivery.

However, as mentioned in the partnership's recent report, “Public AI,” artificial intelligence tools pose significant risks to these efforts. Irresponsible deployments can perpetuate bias, inaccuracy and opaque decisions.

Since the publication of “In the Public AI,” the introduction of powerful, unprecedented generative AI tools has made it even more important for civil servants to understand the trade-offs between AI promises and risk. During this period of uncertainty, responsible use principles derived from previous generations of AI can help eradicate and guide conversations about the possibility of generative AI use.

What is Generated AI? How may it affect public service delivery?

Generation as Category AI is a branch of deep learning artificial intelligence technology that generates new text, images, or other content by replicating the patterns and structure of the data being trained.

Generated AI can provide unprecedented efficiency in the delivery of public services by assuming or supporting the tasks government workers perform. However, errors are common to existing AI systems. Given the wide availability and persuasive output of the generator AI, these tools are extremely dangerous to use without proper understanding.

Apply established best practices

Our report questions for non-technical government leaders distill the lessons learned from previous AI tools into a set of important considerations that civil servants can rely on at each step of the process. Like previous AI tools, agency leaders considering using generative AI tools should fully understand the trade-offs and clarify why generator AI is suitable for your job. Previous iterations of AI deployment revealed important questions to pay attention to at each step of procurement, implementation and evaluation of AI tools to mitigate potential negative impacts.

Similarly, the set of basic AI deployment best practices outlined in Building Blocks for responsible AI usage frameworks is particularly relevant to generating AI.

These practices include maintaining high-quality data, understanding data bias, hiring an experienced workforce, and creating agile government structures that allow for collaboration and educated decision-making. While important for all aspects of agency work, these actions are important in the context of AI, particularly generative AI implementations.

To apply previous iterations of AI in public service delivery, you can read ” to enable you to navigate the current wave of generated AI.With general AI.”