The filing, posted on September 19, is closely redacted and Hive cofounder and CEO Kevin Guo instructed MIT Expertise Evaluate that he couldn’t talk about the main points of the contract, however confirmed it includes use of the corporate’s AI detection algorithms for little one sexual abuse materials (CSAM).
The submitting quotes data from the Nationwide Heart for Lacking and Exploited Youngsters that reported a 1,325% improve in incidents involving generative AI in 2024. “The sheer quantity of digital content material circulating on-line necessitates using automated instruments to course of and analyze knowledge effectively,” the submitting reads.
The primary precedence of kid exploitation investigators is to search out and cease any abuse presently occurring, however the flood of AI-generated CSAM has made it troublesome for investigators to know whether or not photos depict an actual sufferer presently in danger. A software that would efficiently flag actual victims could be a large assist after they attempt to prioritize instances.
Figuring out AI-generated photos “ensures that investigative sources are targeted on instances involving actual victims, maximizing this system’s affect and safeguarding susceptible people,” the submitting reads.
Hive AI provides AI instruments that create movies and pictures, in addition to a variety of content material moderation instruments that may flag violence, spam, and sexual materials and even establish celebrities. In December, MIT Expertise Evaluate reported that the corporate was promoting its deepfake-detection know-how to the US army.
