Artificial Intelligence Working Alone Can't Rescue Government: FDA's Correct Steps and Future Actions
In a significant development, the Food and Drug Administration (FDA) is set to roll out a full generative AI (GenAI) system across all its centers by June 30, marking a new era for government agencies in leveraging advanced technology for efficiency and strategic gains.
The FDA's success with GenAI is not attributed to the best model but rather the extensive work put into organising data, securing access, and choosing high-impact use cases. This approach, which serves as a blueprint for federal agencies, emphasises the importance of proper preparation, security, and return on investment (ROI).
In terms of data preparation, human-centered frameworks and co-design are key. By engaging users early in the process, AI tools can be tailored to fit real workflows, improving adoption and data quality. Investing in AI literacy and continuous learning among staff is also crucial, ensuring they can contribute effectively to data preparation and usage. Infrastructure readiness is another essential factor, with upgrades necessary to handle large-scale GenAI data workloads efficiently.
Security is another critical aspect. Robust security measures, continuous monitoring, and risk management frameworks are essential to protect data privacy, manage risks, and maintain compliance with government standards. Ethical AI usage policies should also be developed to promote responsible use and regular policy updates aligned with evolving technology.
The ROI of GenAI projects is best realised when measurable goals are defined early, including how AI improves productivity, service delivery, and decision-making processes. Tracking user engagement and outcomes is also crucial to optimise AI deployment continuously. Rapid realisation of benefits is expected, with evidence showing that 94% of public sector organisations using GenAI expect ROI within one year.
It's worth noting that GenAI is most effective as a decision support system, not as a replacement for human expertise. If GenAI is deployed without fixing underlying infrastructure and strategy, it may collapse. The FDA's progress with GenAI reflects years of internal work, including data integration, system security, and operational readiness.
In federal environments, GenAI must tailor answers based on who's asking for security and trust reasons. Agencies need built-in, role-based access and row-level data security for GenAI to be safe and scalable. Better data can significantly improve the accuracy of GenAI models, with modern data platforms reducing implementation time from months to weeks and eliminating brittle ETL processes.
Training and trust are as important as technology in the adoption of GenAI. The moment for GenAI adoption in federal agencies has arrived, and the outcome will depend on how agencies proceed. Every agency should follow the FDA's roadmap: start with data organisation, build secure access, empower personnel, measure everything, and remember that GenAI is a tool for transformation, not the mission itself.
By following these guidelines, government agencies can effectively harness generative AI to improve public services while safeguarding data, ensuring security, and maximising return on investment. The FDA's automation of tedious work while keeping scientists empowered and involved serves as a promising example of the potential benefits of GenAI.
The FDA's strategic approach to implementing generative AI (GenAI) across their federal workforce is setting a precedent for other agencies, demonstrating the significance of proper data preparation, security, and return on investment (ROI) in the reimagined workforce. By integrating human-centered frameworks, AI literacy, and continuous learning in their workforce, and ensuring infrastructure readiness, they are making progress in the efficient handling of large-scale GenAI data workloads. The federal workforce, equipped with such technology and artificial-intelligence, can continue to advance and deliver improved public services while maintaining data privacy, managing risks, and maximizing ROI.
