Wired
Facebook, a major SNS company, announced on October 17, local time, that its efforts have halved the number of hate speech posts posted on Facebook and reduced the spread rate to 0.05%.
Facebook publishes hate speech spread rate
According to Facebook, the company operates a system that automatically detects hate speech, and after the content is confirmed by the person in charge, posts are deleted.
According to the company, hate speech was posted on Facebook and the “spread rate” that users can view was 0.05% in the second quarter of 2021 (April to June).
Since the spread rate in the third quarter of 2020 was 0.10% to 0.11%, Facebook claims that the spread rate could be halved in 9 months by operating the system.
It seems that the improvement and expansion of artificial intelligence (AI) used in the system contributed to the reduction of the spread rate of hate speech.
Efforts have been made to delete posts such as hate speech since 2016, and at the beginning, most of the detection of deletion targets relied on reports from users, but now it depends on the system operated by Facebook. Detection has reached 97%.
Is it against leak information?
In early October, a former employee who was a product manager of Facebook leaked internal materials and accused the company of neglecting to protect users in favor of profits, and criticism of Facebook is increasing. ..
Facebook argued in this press release that “leaked internal document data” pointed out that the company’s efforts were inappropriate and concealed, “that’s not true.” Did.
With this announcement, it may be aimed at showing that it is actively preventing the spread of hate speech and promoting the attitude of emphasizing user protection.
Source: Facebook
(seng)
Source: iPhone Mania