Chatroulette is on the rise again – with the help of AI

A decade ago, Chatroulette has been a supernova on the internet, exploding in popularity before collapsing under a torrent of male nudity that rejects users. Now, the app, which randomly pairs strangers for video chats, has a second chance, in part due to a pandemic that has restricted social contact in person, but also due to advances in artificial intelligence that helps filter out the most unacceptable images.

User traffic has tripled since the beginning of the year, reaching 4 million unique monthly visitors, the most since the beginning of 2016, according to Google Analytics. Founder and President Andrey Ternovskiy says the platform provides a refreshing antidote to the diversity and serenity of familiar social echo cameras. On Chatroulette, strangers meet anonymously and do not have to provide their data or delete through advertisements.

A sign of how well Chatroulette has cleaned up its act: an embryonic business of corporate conferences. Bits & Pretzels, a German conference on startups, hosted a three-day Chatroulette event in September, including a Founders Roulette session that matched the participants. “No nudity, but full of surprising conversations,” the conference announced. Another change: women are now 34% of users, up from 11% two years ago.

AI that helped visitors avoid nudity or unwanted masturbation was a good investment, says Ternovskiy. It can also provide lessons for much larger social networks that struggle to moderate content that can deviate from fakes or toxicity. But Ternovskiy still dreams of a platform that creates happy human connections and warns that technology cannot achieve this alone. “I doubt the device will ever be able to predict: is this content desired for my user base?” he says.

A 17-year-old Ternovsky coded and created Chatroulette in November 2009 from his bedroom in Moscow as a way to kill boredom. Three months later, the site attracted 1.2 million daily visitors. Then came the exodus. Ternovskiy has engaged in some unfortunate partnerships with Sean Parker and others in an attempt to maintain relevant Chatroulette. In 2014, it launched a premium offering that associated users based on the desired demographics, which generated some revenue. He has invested some of this money in cryptocurrency projects that have brought in additional profits. Today Chatroulette is based in Zug, Switzerland, an encryption center.

In 2019, Ternovskiy decided to once again give Chatroulette a more respectable business, run by a team of professionals, with less “adult chaos”. The company was incorporated in Switzerland. Ternovskiy hired Andrew Done, an Australian with experience in machine learning, as CTO. Earlier this year, Done became CEO. He was joined by a senior product researcher with a PhD in psychology, a community manager, a talent acquisition manager and several engineers. Then Covid-19 hit and traffic exploded.

The new team leveraged increased traffic to conduct user research and test ways to moderate content, including AI tools from Amazon and Microsoft. Created a filtered channel, now known as Random Chat, designed to exclude nudity, along with an unmodified channel. By delimiting the two channels, Chatroulette hopes to make the filtered feed feel more secure and attract users interested in the human connection. The unfiltered canal remains popular, but its use is dwindling, and Ternovskiy intends to eliminate it by mid-2021.

In June, Chatroulette brought in Hive, a San Francisco-based AI specialist, for a nudity test. Hive software also moderates Reddit content. Executives were quickly impressed by Hive’s accuracy, especially as they did not report users and innocent actions. At the same time, Chatroulette tested the moderation tools from Amazon Rekognition and Microsoft Azure; previously tried Vision AI from Google Cloud.

“Hive is at a level of accuracy that makes it practical to use this technology at scale, which was not possible until now,” says Done. He says Hive is “so accurate that using people in the moderation loop is detrimental to system performance. That is, people make more mistakes than they eliminate. ”

.Source