Study finds AI drafting tools alter user messages on political topics, may affect public opinion

Researchers from Oxford and the University of Potsdam have warned that minor alterations made by artificial‑intelligence tools when redrafting or summarising online content could trigger rapid, widespread shifts in public opinion. Their study finds that AI‑driven platforms, increasingly promoted by tech firms as efficient solutions for handling the daily surge of digital messages, are often embedding partisan slants into the text they process. The biases identified range from distinctly right‑wing to more liberal perspectives, affecting topics as sensitive as abortion, climate change and other political issues.

The analysis examined a broad sample of AI‑generated rewrites across multiple platforms, comparing the original messages with the altered outputs. Researchers observed that even subtle changes—such as word choice, framing or emphasis—can amplify particular viewpoints and diminish opposing arguments. Over time, these cumulative modifications may snowball, influencing how large audiences perceive and discuss contentious subjects, thereby reshaping long‑term public sentiment.

The findings highlight the need for greater transparency and oversight of AI content‑generation tools, especially as they become integral to information dissemination. Policymakers and technology providers are urged to consider safeguards that mitigate partisan distortion, ensuring that AI assistance does not inadvertently steer public discourse.