Facebook executives have long argued that artificial intelligence will solve the company's
longstanding problem of excluding hate speech, overly violent content and underage users from the platform, the Wall Street Journal reported.
Internal documents show that that prospect is far more distant than executives portrayed. Facebook's AI consistently failed to recognize a first-person shooting video, racist remarks or even the difference between a cockfight and a car crash, baffling in-house researchers for weeks.
When it comes to hate speech, Facebook employees expect the company to remove only a small percentage of offending content, "roughly in the single digits," according to the internal document.
IT Home has learned that when Facebook's algorithms are unable to determine whether content is illegal and needs to be removed, the platform shows users less often, but accounts that post content are not penalized for doing so.
About Manly Battery
A leading battery manufacturer mainly offering rechargeable Lithium batteries over 12 years ,widly used for artificial intelligence project ,if there is any project nee to evluate ,pls feel free to send email
to info@manlybatteries.com
