How persuasive is AI-generated propaganda?

Can large language models, a form of artificial intelligence (AI), generate persuasive propaganda?