JMIR Publications

ABSTRACT

Background:

Communication campaigns utilizing social media can raise public awareness, but they are difficult to sustain. One barrier is the need to constantly generate and post novel, yet on-topic messages, which creates a resource-intensive bottleneck.

Objective:

Here, we harness the latest advances in artificial intelligence (AI) to build a system that can generate a large number of candidate messages, which could be used for a campaign. The topic of folic acid, a B-vitamin that helps prevent major birth defects, serves as an example, but the system can work with other topics as well.

Methods:

We used the Generative-Pre-trained-Transformer-2 (GPT2) architecture, a machine learning model trained on a large natural language corpus, and fine tuned it using a dataset of auto-downloaded tweets about #folicacid. The fine tuned model was then used as a message engine, that is to create new messages about this topic. We carried out an online study to gauge how human raters evaluate the AI-generated tweet messages compared to original, human-crafted messages.

Results:

We find that the Folic Acid Message Engine can easily create several hundreds of new messages that appear natural to humans. Online raters evaluated the clarity and quality of a selected sample AI-generated tweets as on par with human-generated ones. Overall, these results show that it is feasible to use such a message engine to suggest messages for online campaigns.

Conclusions:

The message engine can serve as a starting point for more sophisticated AI-guided message creation systems for health communication. Beyond the practical potential of such systems for campaigns in the age of social media, they also hold great scientific potential for quantitative analysis of message characteristics that promote successful communication. We discuss future developments and obvious ethical challenges that need to be addressed as AI technologies for health persuasion enter the stage.