The campaign, dubbed Project Artemis, got under way in November 2018 during a Microsoft 360 Cross-Industry Hackathon that was co-sponsored by WePROTECT Global Alliance and the Child Dignity Alliance. In collaboration with The Meet Group, Roblox, Kik and Thorn, the initiative aims to identify potential instances of online child grooming by applying a technique to historical text-based conversations.
As Microsoft highlights in a recent blog post, the technique evaluates and “rates” conversation characteristics and assigns an overall probability rating to the chat. This rating can then be used to determine if a conversation should be flagged as potentially inappropriate and passed along to human moderators for review. From there, it would be up to the moderators to determine if further action – such as involving law enforcement – is necessary.
“We are making the tool available at this point in time to invite further contributions and engagement from other technology companies and organizations with the goal of continuous improvement and refinement.”
Microsoft said it has used the technique on its Xbox platform “for several years” and is exploring its use in other chat platforms like Skype.
From January 10, companies and services interested in testing out the technique can contact Thorn for licensing and adoption instructions.
Masthead credit: Laptop by GaudiLab