Meta Shifts to AI for Risk Assessment in Product Features
In a bold move, social media giant Meta is considering replacing human reviewers with artificial intelligence (AI) to assess potential risks in its products and features.This shift aims to streamline the evaluation process,allowing the company to reduce the time taken for risk assessments considerably. Internal documents reveal that Meta plans to delegate up to 90% of risk evaluation tasks to AI, particularly focusing on issues related to children and teenagers, including the spread of misinformation and content integrity.
Though, some current and former Meta employees, who wish to remain anonymous, have raised concerns about this approach. They warn that AI may overlook critical risks that humans could easily identify. Traditionally, new features and updates on platforms like Instagram and WhatsApp underwent thorough human review before being released to the public. Yet, reports indicate that in the past two months, Meta has intensified its reliance on AI for these evaluations.Now, before launching any new feature, teams must submit a form to AI, which quickly identifies potential issues or risks. the teams then address these concerns before the feature is finally rolled out to users.
