A client wants an audit of local files across more than 60 workstations. Is there a tool or endpoint agent that can flag likely adult images or videos without making technicians manually review every thumbnail?
What can scan 60-plus workstations for adult content?
Answers
One possible route is to collect files from accessible SMB or NFS shares into Amazon S3 with AWS DataSync, then use Amazon Rekognition for content moderation. That's a custom pipeline, not a ready-made endpoint scanner.
You'll need a way to reach local workstation files, handle files that are locked or unavailable, and decide how to process videos as well as images. Test a small, representative sample first and check how well the classifications match the client's policy. Moving employee files into cloud storage also needs explicit approval and a plan for access and retention.
Reply to this answer
Check whether the client's endpoint security product can inventory or scan files remotely. You'd still need to verify that it actually classifies image or video content. A file search alone may just give you a very long list of media files.
Ask vendors about local drive coverage, offline machines, video support, resource usage, and what information gets sent off the endpoint. Pilot it before committing to a full scan.
Reply to this answer
If the immediate goal is an inventory, remote administrative shares or PowerShell may save a lot of desk visits. You can search for media extensions remotely, although that won't tell you whether the content is adult and will still leave a review burden.
trolleybell.roughly:
You'll also need the right permissions and network access to each machine. An offline or inaccessible workstation should be reported as not scanned, not clean.
Reply to this answer
Is the client trying to find existing files, or stop future browsing? Those are different jobs. A web filter, such as Zscaler or a URL filtering feature in an endpoint product, may help with the latter, but it won't establish what's already stored on local drives.
Firewall or DNS logs can offer clues about network activity, though they aren't a substitute for inspecting files either. I'd get the scope and policy agreed before treating either kind of signal as a finding.
Reply to this answer
Before picking a tool, agree on what counts as in scope, who can see flagged material, and how findings will be handled. Image classification can miss content or flag harmless material, so build in a limited review process and don't promise a definitive result.
Reply to this answer
Add your answer
Answers are reviewed before publication.
Dovlin+perhaps:
That could reduce thumbnail review, but it won't eliminate it. A flagged result is a lead for an authorized reviewer, not proof of a policy violation.