Social media exposure to opposing views can increase political polarization. A study examined the effects of following Twitter bots that exposed users to messages from opposing political ideologies. Participants were Democrats and Republicans who visited Twitter at least three times a week. One week later, they were randomly assigned to follow a liberal or conservative bot for a month. After the intervention, Republicans who followed a liberal bot became more conservative, while Democrats who followed a conservative bot showed slight liberalization, though not statistically significant. The study found that exposure to opposing views can increase polarization, particularly among Republicans. The findings suggest that social media may exacerbate political polarization by reinforcing existing beliefs through selective exposure. The study highlights the importance of understanding the mechanisms behind political polarization and the role of social media in shaping political attitudes. Despite limitations, the study contributes to the field of computational social science and offers insights into reducing political polarization online. The research used a large field experiment and controlled for various factors, including compliance with the intervention. The results indicate that prolonged exposure to counterattitudinal messages can lead to increased polarization, especially among conservatives. The study underscores the need for further research to understand the causal pathways of backfire effects and to develop effective strategies for reducing political polarization in online environments.Social media exposure to opposing views can increase political polarization. A study examined the effects of following Twitter bots that exposed users to messages from opposing political ideologies. Participants were Democrats and Republicans who visited Twitter at least three times a week. One week later, they were randomly assigned to follow a liberal or conservative bot for a month. After the intervention, Republicans who followed a liberal bot became more conservative, while Democrats who followed a conservative bot showed slight liberalization, though not statistically significant. The study found that exposure to opposing views can increase polarization, particularly among Republicans. The findings suggest that social media may exacerbate political polarization by reinforcing existing beliefs through selective exposure. The study highlights the importance of understanding the mechanisms behind political polarization and the role of social media in shaping political attitudes. Despite limitations, the study contributes to the field of computational social science and offers insights into reducing political polarization online. The research used a large field experiment and controlled for various factors, including compliance with the intervention. The results indicate that prolonged exposure to counterattitudinal messages can lead to increased polarization, especially among conservatives. The study underscores the need for further research to understand the causal pathways of backfire effects and to develop effective strategies for reducing political polarization in online environments.