Microsoft publishes a tool to detect `` predators '' sexually exploiting children from conversation


by

monkeybusiness

Microsoft has been strengthening its efforts on child pornography for some time, and experts who detect child pornography are overburdened , so we have also developed detection software using artificial intelligence (AI) . Microsoft has released a new codename ' Project Artemis ', a technology that detects online 'predators' trying to lure children for sexual purposes.

Microsoft shares new technique to address online grooming of children for sexual purposes-Microsoft on the Issues
https://blogs.microsoft.com/on-the-issues/2020/01/09/artemis-online-grooming-detection/



Project Artemis is a technology that began development in the wake of the hackathon held in November 2018. The engineering is being done by Microsoft, as well as social entertainment services The Meet Group , game development platform Roblox , messenger app Kik , and nonprofit Thorn , under the umbrella of Professor Honey Farid, who specializes in digital image analysis. He has been cooperating for 14 months.

The newly developed technology, based on Microsoft patents, applies text-based conversations, assessing the characteristics of the conversation and assigning an overall probability rating. Each company that introduces the new technology uses this rating to flag user conversations, and any conversations that detect problems are sent to a human moderator. The human moderator then determines if there are any threats, such as sexual exploitation.


by bialasiewicz

Microsoft has announced that this technology will be available to businesses free of charge through Thorn from January 11, 2020.

Courtney Gregoire, Microsoft's Chief Digital Security Officer who oversaw the project, said, `` Project Artemis is a big step, but it is not a panacea '', but Microsoft stops moving just because the problem is complicated and messy. Instead, he is committed to making continuous improvements by making the tools open.

in Software, Posted by darkhorse_log