Meta Plans Staff Cuts as It Expands Use of AI for Content and Ad Review

Meta is accelerating its plan to replace human moderators with AI technology for content and advertisement review on platforms like Facebook and Instagram, aiming to reduce staff by the end of this year. The company targets conducting 90 percent of these tasks using AI tools by the end of 2026, raising concerns about security risks and technical vulnerabilities.

Following a recent security breach where hackers manipulated Meta’s AI support bot to gain control of approximately 20,000 Instagram accounts, serious questions have emerged regarding the sensitivity and reliability of AI-driven systems.

As reported on 15 Ashad from Kathmandu, Meta has intensified efforts to displace employees by deploying large language model (LLM) AI systems for reviewing most content and advertisements on its platforms within the current year. According to the Financial Times, Meta is moving forward with plans to replace ‘human moderators’ with AI models. Currently, about 50 percent of such tasks are handled by AI, with the company aiming to increase this figure to 90 percent by the end of 2026. This shift means that nearly all decisions regarding posts and ads on Facebook and Instagram will be made by AI.

This decision comes amid recent technical weaknesses exposed on Instagram, where hackers exploited a vulnerability to gain unauthorized access to roughly 20,000 accounts. The attackers deceived Meta’s AI support bot into sending account authentication codes to alternate email addresses, effectively taking over accounts without requiring any coding or technical skills. Meta has reportedly addressed this vulnerability, but the incident highlights the sensitivity of AI systems. Because users interact with AI using natural language and can ask questions in countless ways, such security weaknesses are likely to persist. Even when AI models are programmed to block certain types of queries, they can be circumvented by phrasing questions differently—a problem seen across multiple AI platforms.

Consequently, entrusting highly sensitive tasks to AI agents increases the risk that users may manipulate these systems to gain unauthorized account access. Nevertheless, Meta considers it necessary to assign these functions to AI. The company has invested billions of dollars in AI development and plans to commercialize the technology to help other firms reduce labor costs.

If Meta fails to demonstrate successful application of AI in its own operations, it could jeopardize the future of its AI projects. Therefore, Meta appears eager to showcase the practical importance of its AI systems by cutting staff and replacing human roles with AI. However, the rapid pace of this transition is also raising significant security concerns.