Technology
Case StudyiQiyi

iQiyi automatically deletes 10% of censorship violating user uploaded content minimising the number of human censors needed through the use of machine learning

The Chinese video platform iQiyi automates censorship duties of user generated content to reduce the number of human censors needed.

Context

"Chinese internet companies and government departments employ several million censors to police content that the ruling Communist party finds objectionable, according to local media. As China’s online population has boomed, human censors have been overwhelmed with an ever-growing wave of online content. Their job is to remove pornographic and violent content, as well as themes deemed “politically sensitive” by the authoritarian government, such as homosexuality. Censorship of material is an increasingly costly, as well as risky, exercise for internet companies in China."

The Project

"iQiyi, a top video-streaming platform, has turned to machine learning to filter content Beijing wants to ban. IQiyi’s algorithms automatically delete about 10 per cent of all content uploaded. The algorithms then grade the remaining videos for their risk of containing illegal material and human censors review potentially problematic videos."

Results

iQiyi''s team of content censors only doubled from 2014-2017, despite a 20-fold increase in videos on the platform in the same amount of time.

Back to Case Studies
AI Daily Brief — leaders actually read it.

Free email — not hiring or booking. Optional BPAI updates for company news. Unsubscribe anytime.

Include

No spam. Unsubscribe anytime. Privacy policy.