Meta will reportedly soon use AI for most product risk assessments instead of human reviewers

According to a report Meta plans to release the task of moving its products away from human reviewers, to accelerate the term, to determine the task of removing material losses from human reviewers. The domestic documents of the publication, AI aims to have Meta risk assessments to be 90 percent, AI, Npr Reports and AI reviews, even young risk and integrity, such as “integrity”, even the violent content, the wrong information and more. Unnamed Current and Old Meta Workers Npr Warned AI can review the serious risks that a human team can determine.

News and new features for meta platforms, including Instagram and WhatsApp, have long been subjected to human reviews before the people are hit, but Meta has been twice the use of AI for the last two months. Now, according to Npr, Product teams must complete a questionnaire about the product and submit it to review it AI systemIn general, the “a sudden decision”, which includes the areas of the established risk. You must then settle any requirements to solve problems before the product is released.

An old meta executive director reported Npr The research reduction means that you have created higher risks. There are fewer external features of product changes before the world causes problems. ” In a statement NprMeta will still hit “human experience” and “low-risk decisions” to evaluate “novel and complex issues” and “low-risk decisions” will release the EU to AI. Read the full report .

META is a few days of released – since today and at the beginning of this year. After the changes in the report, the amount of content shown below has decreased surprisingly. However, there was a small rise in violence and aggression, but also in severe and graphic content.