Every April, organizations across industries celebrate Records Management Month, as it is a time dedicated to recognizing the critical role that records and information professionals play in organizing, preserving, and protecting information. In today’s digital-first world, effective records management is no longer optional; it is essential. Records Management Month also offers an opportunity to reflect on one of the most significant forces shaping the profession today: artificial intelligence (AI). The evolution of technology is moving at lightning speed, as government agencies are trying to grasp how records management is imperative to prioritize as the digital world progresses.
According to the Virginia Public Records Act (VPRA), a record accounts for a business transaction and/or activity and is defined by its context and not format. As we also celebrate the 50th anniversary of the VPRA, we must acknowledge how the evolution of technology, including AI, was not particularly considered. Government agencies are currently utilizing AI platforms such as Copilot and ChatGPT to automate administrative tasks, such as document review and data entry, assist with routine inquiries through AI-powered chatbots, and analyze financial data to identify suspicious activity through AI algorithms.
Adaptation to technological evolution is not foreign, as agencies have transitioned into the management of official records in electronic and digital formats such as databases, PDFs, and images, which can enhance operational efficiency and improve data security; the integration of AI could be viewed in the same way. On the contrary, the concern lies in ensuring that business activities (records) generated by AI are captured, retained, and legally dispositioned according to the Library of Virginia’s retention and disposition schedules. The guidance of applying retention and ensuring that AI-generated records are carried out through their lifecycle seems simple, but Library of Virginia records analysts are learning more of the complexities of AI and its impact on effective records management.
Library of Virginia records analysts have experienced an increase in inquiries from agencies asking how AI-generated records should be retained, especially as the retention schedules are written to acknowledge the context of a record and not its format. It is simple to respond with, “If it pertains to business activity, it should be retained,” but AI is a bit different, which has urged analysts to evaluate how AI is used to create records. Library of Virginia analysts are collaborating with the government agencies’ records management subject matter experts to assess how the AI results are generated, which involves evaluating the prompts that a user asks of an AI system. At this very moment, the prompts used are identified as a business activity, and this is where records management comes into play. Also, this is where Library of Virginia analysts begin to ask a plethora of questions that some agencies may not be able to confidently answer: How are these prompts being managed? Can IT professionals capture the logs of AI prompts and results? Have government agencies outlined the use of AI systems within their records management policies?
This leads to agencies understanding how it is imperative to have a grip on the use of AI for the protection of the agency and to maintain trust with citizens of the Commonwealth. There has been a slow rise in citizens’ submissions of Freedom of Information Act (FOIA) requests for AI prompts used by government personnel. As such requests continue to populate within the Commonwealth, government agencies are starting to become aware that their use of AI is considered a record. The major concern associated with this is whether the agency has reinforced which AI platforms are and are not permissible to use for business activity. To offset this issue, agencies are recommended to develop policies outlining the use of AI systems for personnel to understand. Government agencies are also encouraged to implement formal agreements with AI providers rather than using public platforms to ensure data privacy, intellectual property protection, and legal compliance.
The VPRA synthesizes that agency records must contain information needed to protect the interests of the Commonwealth. The integrity of a record is critical in the world of records management, including when a record is digitized, converted, or migrated from one system to another. The same thought is expected of AI-generated records. Without caution, AI can cause a loss of context and quality as automated classification may not accurately identify or categorize records. Additionally, there are major risks of incorrect information, also known as hallucinated information, which can be added to records, resulting in a lack of accountability. As AI continues to become present within government agencies to streamline processes, potentially reduce backlogs, and enhance decision-making, it is important to prioritize the component of records management to account for business activity.
As the evolution of technology is embraced and infused into the operations of agencies, records management should be an important component within AI integrations. In celebration of Records Management Month, government agencies’ records information professionals, leadership, and information technology professionals are encouraged to communicate about and analyze how their implementation of AI is aligned with the requirements of the Virginia Public Records Act and to ensure records are efficiently captured to account for business transactions. AI is expected to progress as the digital world continues to evolve, and agencies should be proactive, rather than reactive, to sustain the fiscal, administrative, legal, and historical value of their records.
