a spammer is motivated by self-enrichment

AI agents are grounded on the user provided direction. If it doesn't work well then it naturally loses trust.