US GOVERNMENT TURNS TO AI FOR DOCUMENT REDACTION

Federal agencies plan to use AI tools to identify and remove classified or sensitive material from documents before public release. The technology promises faster processing of the thousands of FOIA requests received annually. However, privacy advocates and transparency experts warn the approach may be counterproductive. AI redaction systems can be overly aggressive, flagging information that doesn't actually require withholding. This could result in citizens and journalists receiving less disclosure than legally required. The effectiveness of AI redaction depends heavily on how agencies train and configure the systems. Poorly implemented AI could become a bottleneck rather than a solution, potentially denying legitimate access to public records. Agencies have not yet disclosed which specific AI tools they plan to use or detailed testing results. The rollout timing and scope across different government departments remain unclear.