Position: Political Neutrality in AI Is Impossible — But Here Is How to Approximate It

Jillian Fisher, Ruth Elisabeth Appel, Chan Young Park, Yujin Potter, Liwei Jiang, Taylor Sorensen, Shangbin Feng, Yulia Tsvetkov, Margaret Roberts, Jennifer Pan, Dawn Song, Yejin Choi
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:81315-81374, 2025.

Abstract

AI systems often exhibit political bias, influencing users’ opinions and decisions. While political neutrality—defined as the absence of bias—is often seen as an ideal solution for fairness and safety, this position paper argues that true political neutrality is neither feasible nor universally desirable due to its subjective nature and the biases inherent in AI training data, algorithms, and user interactions. However, inspired by Joseph Raz’s philosophical insight that "neutrality [...] can be a matter of degree" (Raz, 1986), we argue that striving for some neutrality remains essential for promoting balanced AI interactions and mitigating user manipulation. Therefore, we use the term "approximation" of political neutrality to shift the focus from unattainable absolutes to achievable, practical proxies. We propose eight techniques for approximating neutrality across three levels of conceptualizing AI, examining their trade-offs and implementation strategies. In addition, we explore two concrete applications of these approximations to illustrate their practicality. Finally, we assess our framework on current large language models (LLMs) at the output level, providing a demonstration of how it can be evaluated. This work seeks to advance nuanced discussions of political neutrality in AI and promote the development of responsible, aligned language models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-fisher25a, title = {Position: Political Neutrality in {AI} Is Impossible — But Here Is How to Approximate It}, author = {Fisher, Jillian and Appel, Ruth Elisabeth and Park, Chan Young and Potter, Yujin and Jiang, Liwei and Sorensen, Taylor and Feng, Shangbin and Tsvetkov, Yulia and Roberts, Margaret and Pan, Jennifer and Song, Dawn and Choi, Yejin}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {81315--81374}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/fisher25a/fisher25a.pdf}, url = {https://proceedings.mlr.press/v267/fisher25a.html}, abstract = {AI systems often exhibit political bias, influencing users’ opinions and decisions. While political neutrality—defined as the absence of bias—is often seen as an ideal solution for fairness and safety, this position paper argues that true political neutrality is neither feasible nor universally desirable due to its subjective nature and the biases inherent in AI training data, algorithms, and user interactions. However, inspired by Joseph Raz’s philosophical insight that "neutrality [...] can be a matter of degree" (Raz, 1986), we argue that striving for some neutrality remains essential for promoting balanced AI interactions and mitigating user manipulation. Therefore, we use the term "approximation" of political neutrality to shift the focus from unattainable absolutes to achievable, practical proxies. We propose eight techniques for approximating neutrality across three levels of conceptualizing AI, examining their trade-offs and implementation strategies. In addition, we explore two concrete applications of these approximations to illustrate their practicality. Finally, we assess our framework on current large language models (LLMs) at the output level, providing a demonstration of how it can be evaluated. This work seeks to advance nuanced discussions of political neutrality in AI and promote the development of responsible, aligned language models.} }
Endnote
%0 Conference Paper %T Position: Political Neutrality in AI Is Impossible — But Here Is How to Approximate It %A Jillian Fisher %A Ruth Elisabeth Appel %A Chan Young Park %A Yujin Potter %A Liwei Jiang %A Taylor Sorensen %A Shangbin Feng %A Yulia Tsvetkov %A Margaret Roberts %A Jennifer Pan %A Dawn Song %A Yejin Choi %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-fisher25a %I PMLR %P 81315--81374 %U https://proceedings.mlr.press/v267/fisher25a.html %V 267 %X AI systems often exhibit political bias, influencing users’ opinions and decisions. While political neutrality—defined as the absence of bias—is often seen as an ideal solution for fairness and safety, this position paper argues that true political neutrality is neither feasible nor universally desirable due to its subjective nature and the biases inherent in AI training data, algorithms, and user interactions. However, inspired by Joseph Raz’s philosophical insight that "neutrality [...] can be a matter of degree" (Raz, 1986), we argue that striving for some neutrality remains essential for promoting balanced AI interactions and mitigating user manipulation. Therefore, we use the term "approximation" of political neutrality to shift the focus from unattainable absolutes to achievable, practical proxies. We propose eight techniques for approximating neutrality across three levels of conceptualizing AI, examining their trade-offs and implementation strategies. In addition, we explore two concrete applications of these approximations to illustrate their practicality. Finally, we assess our framework on current large language models (LLMs) at the output level, providing a demonstration of how it can be evaluated. This work seeks to advance nuanced discussions of political neutrality in AI and promote the development of responsible, aligned language models.
APA
Fisher, J., Appel, R.E., Park, C.Y., Potter, Y., Jiang, L., Sorensen, T., Feng, S., Tsvetkov, Y., Roberts, M., Pan, J., Song, D. & Choi, Y.. (2025). Position: Political Neutrality in AI Is Impossible — But Here Is How to Approximate It. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:81315-81374 Available from https://proceedings.mlr.press/v267/fisher25a.html.

Related Material