Article reader Listen + reading controls
Article reader
Preparing the reader…
Reading settings
DHS’s Initial AI Inventory: Verification Issues Highlighted by Watchdog Report According to a report from the Government Accountability Office (GAO), the Department of Homeland Security (DHS) did not properly certify whether the cybersecurity use cases listed in its AI inventory were genuine examples of artificial intelligence. One specific use case, Automated Scoring and Feedback (AS&F), was incorrectly characterized as AI when it was actually a predictive model for sharing cyber threat information. This raises concerns about the accuracy and reliability of DHS’s AI Use Case Inventory. The GAO faulted DHS for not fully implementing the AI Accountability Framework and recommended expanding the AI review process and ensuring the accuracy of inventory submissions. DHS has agreed to all eight recommendations proposed by the GAO, aiming to promote responsible and accountable use of AI and building public trust in the agency’s AI capabilities.¶
DHS’s Initial AI Inventory: Verification Issues Highlighted by Watchdog Report¶
A recent investigation has raised questions regarding the Department of Homeland Security (DHS)’s handling of Artificial Intelligence (AI) use case submissions. According to a Government Accountability Office (GAO) report, DHS was not confirming the authenticity of reported AI applications.
Questions Over DHS’s AI Inventory¶
Apparently, the DHS’s AI inventory, created to address a 2020 executive order regarding AI in federal government, included cases that were not genuine AI examples. This has raised concerns about the overall trustworthiness of DHS’s AI use case inventory.
Specifics of the Report¶
According to the GAO report, one of the two cybersecurity-focused cases listed in the initial inventory was incorrectly characterized as an AI application. This case, called Automated Scoring and Feedback (AS&F), was actually a predictive model meant to share cyber threat information.
Perceived Flaw in DHS’s AI Strategy¶
The DHS has a reviewing process wherein use cases are scrutinized before adding them to the AI Inventory. However, the GAO report notes that DHS does not confirm whether the listed uses are appropriately characterized as AI. The GAO asserts that until DHS expands its reviewing process to include such determinations, accurate use case reporting cannot be ensured.
Conclusion¶
The report from GAO shares insights into potential inaccuracies in DHS’s AI Inventory, thereby sharpening the focus on the necessary procedures to improve AI use case reporting accuracy. For a comprehensive understanding of this report, visit the original FedScoop article .