The rise of AI and its applications (generative imagery, animation, text and much more), the expected productivity dividends and more is already covered elsewhere. But as this technology enhances efficiency, it comes with risks which need to be considered.
We’ve already explored AI and Information Management before, but recent news has provoked new discussion about AI, records and data security; particularly with the recent news that the Department for Homeland Security (DHS) in the U.S. will be piloting AI technology in a variety of key areas (including records). So we’ve offered a rundown of the intersection between this exciting new technology, as well as some of the data security concerns.
So how does AI help with records management anyway?
If you’re not already familiar with the potential application of AI in terms of AI and Information Management, we’ve listed a few of the tasks it is already being used for below:
- Automating data capture: AI can automatically extract and classify information from various sources, reducing manual entry errors. For example, AI tools can scan documents and emails to capture relevant data points for compliance or archival purposes without human intervention.
- Improving data retrieval: Advanced search capabilities enable users to retrieve documents using natural language queries. For instance, in healthcare, clinicians can use prompts like “Show all diabetic patient records from 2023” to quickly access specific data.
- Ensuring consistent data entry: AI can standardize the information entry process across platforms, making sure that all data adheres to the same formats and standards, which is crucial for large organizations with multiple data entry points.
- Enabling smarter categorization and filing: AI systems can analyze the content of documents and categorize them accordingly, such as automatically tagging all legal documents with specific compliance needs or privacy levels.
Generative AI in government: DHS’s approach
As of April 2024, the Department for Homeland Security (DHS) in the United States is exploring generative AI to enhance its operations.
A prime example includes the Federal Emergency Management Agency’s pilot project, which explores hazard mitigation planning and training projects with U.S Citizen and Immigration Services. Such initiatives of generative AI illustrate the potential of AI to streamline government processes.
A significant amendment in OpenAI’s terms of service, as noted by FedScoop, highlights that outputs from AI models are considered federal records under the Freedom of Information Act (FOIA). These alterations emphasize the need for clear guidelines on how AI interactions are managed as records.
This is one of the first times we’ve seen concerns over the intersection between AI-powered records management and security raised by public figures, so expect more conversation about this over the coming year. Although it’s just pilot projects and a broad roadmap for now, we expect to see uptake of generative AI become more entrenched over the course of 2024, across government organizations and private enterprise. The DHS’s approach of piloting in certain areas, alongside a huge emphasis on educating employees about AI processes, as well as ensuring rigorous security will become a model to follow.
What are the legal considerations of AI from a records perspective?
The integration of AI tools like ChatGPT in government operations necessitates reevaluating records management frameworks to ensure transparency and compliance. The accessibility and searchability of records for FOIA requests are critical here.
Enid Zhou from the Electronic Privacy Information Center (EPIC) in Washington DC, raised concerns about whether agencies can effectively access and search these AI-integrated records systems. Furthermore, there’s a risk of misusing exemptions, such as the deliberative process privilege, to withhold AI-generated content from public scrutiny, potentially obscuring transparency.
Explainable AI and Enhancing Transparency
Explainable AI (XAI) could potentially address these transparency issues by making the processes behind AI decisions more understandable. As described by the World Conference on XAI, this approach aims to detail the mechanics of AI operations – a sort of revealing of the black box, providing insights into how decisions are made or data is processed. An emphasis on transparency that will doubtlessly become more important as uptake of this technology increases.
This level of clarity is vital for ensuring that AI-enhanced records management complies with legal standards while remaining accessible for audits. For instance, if an AI system decides to redact certain information from a public record, XAI could provide a clear rationale based on specific legal exemptions, thereby maintaining a transparent audit trail.
The degree to which XAI itself could be formalized and regulated isn’t entirely clear yet, but as with government records and AI, this is very much a work in progress.
Generative AI in healthcare records: Epic’s integration
In healthcare, EPIC’s integration of AI into its electronic health records system showcases how AI can transform records management. EPIC’s collaboration with Microsoft and Nuance has led to AI and ambient technologies that automate clinical documentation and response drafting. This not only enhances clinical efficiency but also improves patient interactions. For example, EPIC’s ambient AI technology can listen to a patient consultation and automatically draft visit notes, which are then reviewed by the clinician, significantly reducing the time and effort to manually transcript notes.
The adoption of AI tools across various sectors signifies a transformative shift in records management practices. For records management professionals, the challenge is to leverage AI to enhance efficiency and compliance without compromising the integrity and accessibility of records. Navigating the complexities of integrating AI into existing legal frameworks while maintaining reliable record-keeping systems is essential.
As AI technologies continue to evolve, their potential to revolutionize records management is now becoming more apparent: Whether enhancing transparency in government or improving efficiency in healthcare, AI-driven tools offer a promising future for more responsive and effective records management systems. To companies and organizations, this means better productivity, but the benefits for the average person should not be lost too: the ability to retrieve digital records on demand, without knowledge of complicated records management systems could be a lifesaver for everyone from patients to a family purchasing a new home.
It need not be stated enough: Ensuring these technologies are implemented in ways that uphold good governance and public accountability will be crucial.
Crown Records Management has been one of the world’s leading records and information management firms for well over 30 years. If you’d like to talk about starting your own digital journey, get in touch with us today.
