Facebook has long touted its artificial intelligence technology as the ultimate tool to eliminate hate speech, spam, offensive content, and more from its platform. However, according to internal memos from 2019, at least some of the company’s own engineers expressed doubt that AI will ever be able to keep up with the scale of the problem.

“The problem is that we do not and possibly never will have a model that captures even a majority of integrity harms, particularly in sensitive areas,” a senior engineer and research wrote in a 2019 internal memo. “Recent estimates suggest that unless there is a major change in strategy, it will be very difficult to improve this beyond 10-20% in the short-medium term.”

To make matters worse, internal Facebook documents also reveal that its AI system often misidentifies content that human moderators easily catch. For instance, the moderation tool misidentified a cockfighting video as a car crash, while a livestreamed mass shooting was flagged as a trip through a car wash. Of course, some mistakes should be expected on a case-by-case basis. But if Facebook wants to become as reliant on AI as it says, it will have to do more to prevent these embarrassing slip ups.

Recommended Resources

The Choice of Tech Experts Worldwide. Try 90 days free of Bitdefender and experience the highest level of digital safety.

Surf the web truly incognito. Try Bitdefender Premium VPN, the ultra-fast VPN that keeps your online identity and activities safe from hackers, ISPs and snoops.

System Mechanic 14 – Make your computer run like new. Winner of 200+ Editor’s Choice awards!