Meta is reportedly planning a major shift in its internal review process by allowing artificial intelligence (AI) to manage a majority of risk assessments for new features and product updates across its platforms.
According to an NPR report, the tech giant could automate up to 90 percent of these evaluations, which were previously conducted by human experts.
This potential transition would apply to risk reviews for platforms like Facebook, Instagram, WhatsApp, and Threads.
Until now, all updates—ranging from new algorithms to safety tools—had to undergo assessments by human reviewers. These assessments, known as privacy and integrity reviews, analyse whether a new feature could negatively affect users, compromise privacy, harm minors, or contribute to misinformation or harmful content.
By leveraging AI, Meta aims to speed up the development and launch of new features.
According to the report, product teams would fill out a digital questionnaire about a feature, after which the AI would either give instant approval or generate a list of necessary conditions to be met. The team would then have to confirm that these conditions have been satisfied before proceeding with the launch.
Meta reportedly believes that automating this process will significantly accelerate how quickly updates are released. However, this approach has sparked concerns among both current and former employees. Some worry that eliminating human oversight in high-risk areas could compromise safety and accountability.
In response to NPR, Meta stated that human reviewers are still involved in “novel and complex issues,” and that AI is currently used only for low-risk decisions. Yet, internal documents reportedly indicate that AI may soon be used in more critical areas like AI safety, youth risk, and integrity reviews—categories that include violent content and misinformation.
One unnamed Meta employee who has worked with the product risk team told NPR that the automation rollout began in April and continued through May. “I think it's fairly irresponsible given the intention of why we exist. We provide the human perspective of how things can go wrong,” the employee said.
This development comes shortly after Meta published its Integrity Reports for Q1 2025.
In the report, the company noted, “We are beginning to see LLMs operating beyond that of human performance for select policy areas.” Meta added that AI is already being used to remove content from review queues when the system is “highly confident” that no policy violation has occurred. “This frees up capacity for our reviewers allowing them to prioritise their expertise on content that's more likely to violate,” the report said.