The Finnish startup Utopia Analytics trains machine intelligence to do content moderation on social media platforms, discussion forums, and e-commerce sites (Algorithm Watch, 2019). The firm hopes to “bring democracy back into social media” by encouraging safe community dialogue (ibid). Utopia Analytics sells an AI (specifically, a neural network) that is first trained by humans from a body of sample content and then is left to moderate on its own. Human moderators retain the ability to make decisions in complicated and controversial scenarios.
Despite the service’s implementation on various sites like Swiss peer-to-peer sales platform tutti.ch and Finnish public forum Suomi24, Mark Zuckerberg has denounced Utopia Analytic’s technology as primitive and inaccurate. He claims that automated content moderation is at least 5-10 years away due to the significant progress needed to be made before for machine intelligence can comprehend the subtleties and situational context inherent to cracking human language (ibid).