For text polls (and word cloud as well as Q&A comments), an automatic profanity filter is applied.
If a participant sends in abusive language, it is automatically blocked from appearing on displays in the present view and on the participant devices.
For text polls, if a response contains an abusive word the entire message will be filtered out from being displayed.
This profanity filter can however be turned off. Administrators of Enterprise and Institution accounts can customize their profanity list. Read this article.