Technology giant Microsoft announced that it has developed an artificial intelligence (AI)-based tool for detecting sexual predators attempting to groom children online.

The initiative is dubbed Project Artemis in a nod to Artemis from Greek mythology, who was protector and patron of young girls. The AI tool automatically scans online chat content for textual signs of an adult attempting to groom or exploit a child for nefarious purposes.

Although the tech company offered few details surrounding how the AI would work so as to potentially prevent predators from finding a loophole to undermine the technology, they did reveal that the AI scans historical messages for word patterns and characteristics suggestive of grooming and exploitation. The AI assigns content from each chat session a probability score that, when high enough, triggers a human moderator review. If the human moderator reaches the same conclusion, law enforcement is then reportedly notified.

Developed in conjunction with The Meet Group, Roblox, Kik and Thorn, the technology is already being used on Microsoft’s Skype and Xbox platforms. The company intends to make the text scanning AI available at no cost to other companies that host online chats as well, but warns that the technology has not been perfected yet as AI continues to struggle with context within language. As such, some flagged content may render false positives.

In addition to potentially preventing the exploitation of children, the tool could also be valuable for law enforcement and related agencies, offering them insight into how sexual predators operate online.

Microsoft has made the technology available through Thorn, a non-profit organization that combats child sex abuse.