Addressing misinformation in the digital age poses a complex challenge with significant societal implications, as it continues to influence critical events such as elections and pandemic responses. Despite extensive research, misinformation remains a pressing threat, largely because current mitigation strategies emphasize detection accuracy without fully addressing the social and behavioral perspectives of users. This paper proposes a novel approach, treating information as a social norm and tracking its evolution at the individual user level. By targeting specific users who are more susceptible to adopting certain norms, this paper offers a more effective approach to combating misinformation. To accomplish this, a temporal graph neural network is trained to predict the tipping points of individual users, equipped with projection operators capable of forecasting user behavior in real time. To our knowledge, this paper represents the first successful attempt to predict individual tipping points in a real-world setting.