Are LLMs More Persuasive than Humans?

Author(s): MKWriteshere

Originally published on Towards AI.

Why large language models just proved they can change human minds better than actual humans

Imagine walking into a car dealership where one salesperson is human and the other is an AI chatbot. Both want to sell you a car and both get bonuses for closing deals. After an hour of conversation, you’re more likely to buy from the AI.

The article discusses a groundbreaking study which reveals that large language models (LLMs), specifically Claude Sonnet 3.5, are significantly more persuasive than humans in both truthful and deceptive contexts. Through interactive experiments involving 1,242 participants, the study demonstrated that LLMs not only influenced decisions effectively but also raised ethical concerns regarding AI’s potential for manipulation in a variety of fields, such as sales and misinformation. The findings stress the necessity of developing guidelines to manage the persuasive capabilities of AI, promoting critical thinking and AI literacy among the public to guard against potential deceptive influences.

Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI