Kevin Guo founded Hive as a consumer app in 2014 and later donated the corporate to sell its internal content AI software to customers.
Jamel Topin for Forbes
AGenerated content from illegal child sexual abuse images to misleading political deepfakes is infecting the web, and Hive CEO Kevin Guo believes his company’s content mastery systems are a “modern-day antivirus.”
Hive systems are utilized by social media platforms corresponding to Reddit and Bluesky. Now the startup will probably be higher equipped to discover and take away underage sexual abuse material from its clients’ web sites because of a brand new partnership with the Internet Watch Foundation (IWF), a U.K.-based nonprofit, Hive said Thursday.
Hive will integrate the organization’s datasets, which contain a repeatedly updated list of roughly 8,000 web sites which have host images verified by each real-world and AI-generated CSAM, into its models. The dataset also comprises an inventory of unique phrases and cryptic keywords utilized by criminals to cover CSAM and Rock moderation. As a part of the deal, Hive’s customers can gain access to IWF’s “hashes” – digital fingerprints of tens of millions of known CSAM images and videos. The partnership builds on the Hive 2024 collaboration with the 2024 collaboration mandrela national nonprofit organization that’s constructing CSAM detection technology to expand its reach and flush out violators on its clients’ platforms.
The latest partnership goals to stem the tide of kid sexual abuse images, of which perpetrators created tens of 1000’s in 2024. Generative AI tools have only made it easier to provide illegal images. In 2023, IWF flagged over 275,000 web sites with CSAM to law enforcement agencies, a record number for the organization.
“CSAM was previously somewhat difficult to obtain. The content is not that common,” Guo said Forbes. “These AI generation engines, the image and video both, are unlocking a whole different world here with the explosion of content.”
Hive was founded in 2014 as a social media app and in 2017 planned to sell its internal moderation tools to corporations as a substitute. In addition to detecting toxic content, the corporate’s AI models can even discover logos, recognize tens of 1000’s of celebrities, and share copies of movies and tv shows online. The San Francisco-based company has raised $120 million in enterprise capital from General Catalyst and is valued at $2 billion in 2021.
The wide prevalence of AI-generated content has translated into business growth for Hive, where revenue has multiplied 30 times since 2020 (Guo declined to disclose current numbers). The company processes 10 billion pieces of content every month from its 400 customers, including social media platforms corresponding to streaming platform Kick, which has around 50 million users, and the Pentagon. Last month Hive secured a $2.4 million Contracted with the US Department of Defense to be sure that the audio, video and text-based content provided by employees from various sources is real and trustworthy.
There’s also increasing interest from document verification corporations and insurance firms, which have reported an influx of pretend claims “where people in a car have taken a picture of their bumper and the AI creates a scratch,” Guo said.
Recently, alternative platforms corresponding to Clapper and Favorites have signed up for Hive’s automated content moderation systems to arrange for an impression of “Tiktok refugees,” Guo said. “They have also proactively come on board with our CSAM offerings because they are very fearful of becoming a high-profile issue.”
Guo will not be concerned about President Donald Trump’s hands-off approach to regulating AI. Although he rescinded the Biden Administration’s executive order on AI, which outlines measures to deal with deepfakes, online safety for youngsters is a “pretty bipartisan issue,” Guo said. “We don’t think that part is going to go away.”
More from Forbes
