Story perspectives

AI Drafting Tools Inject Partisan Bias, EU Rules Lag

7/7/2026

Summary

Story summary

Internet Institute and Hasso Plattner Institute researchers reported on 2026-07-06 that AI drafting tools rewrote posts with partisan slants.
Meta, Google, Alibaba’s Qwen and Mistral added liberal bias to climate and gender posts while xAI’s Grok added opposite bias to abortion and religious statements.
Researchers warned that nudges can spread across millions of interactions, shifting public opinion, and that the EU AI Act and Digital Services Act leave the issue unaddressed.

1 of 1

Story summary

Internet Institute and Hasso Plattner Institute researchers reported on 2026-07-06 that AI drafting tools rewrote posts with partisan slants.
Meta, Google, Alibaba’s Qwen and Mistral added liberal bias to climate and gender posts while xAI’s Grok added opposite bias to abortion and religious statements.
Researchers warned that nudges can spread across millions of interactions, shifting public opinion, and that the EU AI Act and Digital Services Act leave the issue unaddressed.