Co-Writing with Opinionated Language Models Affects Users’ Views

If large language models like GPT-3 preferably produce a particular point of, they may influence people's opinions on an unknown scale. This study whether a language-model-powered writing assistant that generates opinions more often than others impacts what users write - and what they. In an online experiment, we asked participants (N=1, 506) to write a post whether social media is good for society. Treatment group used a language-model-powered writing assistant configured to that social media is good or bad for society. Participants then completed social media attitude survey, and independent judges (N=500) evaluated the expressed in their writing. Using the opinionated language model the opinions expressed in participants' writing and shifted their in the subsequent attitude survey. We discuss the wider implications our results and argue that the opinions built into AI language technologies to be monitored and engineered more carefully.