<p>The rapid rise of social media platforms has led to an increased reliance on automated content moderation systems. Although these systems are designed to detect and remove malicious content, they often lead to the censorship of legiti- mate content, including news articles, user comments, and even virtual meeting discussions, due to inherent biases or errors in the detection algorithm. This research aims to develop an AI platform that can effectively distinguish between legitimate and malicious content and ensure that lawful expression is not unfairly suppressed. The proposed platform uses advanced machine learning techniques to assess the context and intent of content such as news reports, social media posts and comments in real-time before flagging it for moderation. By analysing a large dataset of tagged posts, meetings, and news articles, we train a robust model that reduces false positives while maintaining high accuracy in identifying mali- cious material. Our findings show that the platform significantly improves the fairness and transparency of content moderation systems and ensures freedom of expression in online discussions and information sharing. The implications of this work are significant as it can be integrated into existing moderation frameworks to prevent undue censorship and promote a more balanced digital ecosystem.</p>