Facebook’s commitment to countering deemed offensive content (nudity, violence, etc.) Is well known . From today the company Zuckerberg has among its priorities the creation of a artificial intelligence system capable of detecting independently considered inappropriate in live video.
The live streaming is now one of the most used features from Facebook users, and with the increase in the number of users increases the risk of an increase in violations of the policy of the social network. The automated system would do without signaling intervention from the user, often far from objective. Reliability remains the central node of half totally managed by the AI considering the often delicate, and not always objectively classifiable, of certain content. At the time were not provided further technical details: a hypothesis to be considered would be to combine the automated system control of a human case of doubt, however, that monitoring should intervene in real time given the nature of the medium.
Already today the elimination of textual or photographic content often seems dictated by guidelines not entirely clear often even contradictory. One thing is certain: the task of the engineers will not be more simple.