Background and Objective:
Social media were crucial in spreading vaccine hesitancy during the COVID-19 Pandemic. Social media monitoring of vaccine stances can be extremely useful to identify trending topics and adapt communication strategies. The objective of this scoping review is to explore the use of Artificial Intelligence (AI), Such as Natural Language Processing (NLP), in social media monitoring, on the topic of vaccine stance and intent related to COVID-19 vaccination campaigns.

Methods:
The study was conducted in november 2022 on multiple databases, using the pico framework. A search query identifying mesh terms for pubmed was developed and adapted for scopus and web of science databases. The study included only english papers, published during the last 3 years, that specifically addressed ai methodologies, social media and vaccine stance. Search results were uploaded to rayyan, a collaborative web application for systematic reviews, to automatically identify duplicates. All studies identified were screened by five authors using article titles and abstracts. Further steps are ongoing.

Results:
A total of 2,722 results were found on pubmed, 5,230 results on scopus, 3,304 results on web of science. After clearing out the duplicates, 5,674 articles were included for title and abstract screening. Among identified ai methodologies, nlp solutions extracting information from tweets were the most frequently identified. Sentiments identified among different studies were discouraging towards vaccination in 20% of the tweets. Further results will be available after the initial and full-text screening to reach a total of a few tens of articles.

Conclusion:
This review shows the opportunity to apply AI methodologies to social media monitoring regarding COVID19 vaccine stance. Despite the frequent use of NLP, many different machine learning methodologies are reported in the screened articles , which combined with the growing amount of existing social-media, leads to a great heterogeneity in AI-based social media monitoring.