Meta plans to replace humans with artificial intelligence when assessing privacy and risks in its products
Meta plans to significantly change the approach to privacy and risk assessment in its products, gradually replacing most human reviewers with artificial intelligence systems. The company intends to transfer up to 90% of such processes to automated AI models, informs NPR.
In effect, this means that key updates to the Meta algorithms, introduction of new security tools, and changes in content distribution policies will mostly be approved automatically — based on AI system analysis.
While previously every update on Meta platforms, including Instagram and WhatsApp, was subject to mandatory human review, over the past two months the company has significantly stepped up its use of artificial intelligence. Currently, developers fill out a special questionnaire about the product, which AI processes instantly, identifying potential risks.
At the same time, former Meta employees are concerned about the risks associated with automation, as AI systems may miss some important problems that an expert would notice. In response, Meta emphasizes that human factors will continue to be used in cases of complex and high-risk issues, and automated solutions will be used only for standard and minor situations.
These changes coincided with the publication of the company’s quarterly reports on content moderation, which noted a decrease in the number of removed content, but at the same time an increase in the number of incidents of bullying and posts with violent content.
