What is the danger?
AI can hold a conversation, learn who you are and shape its arguments to fit you personally. Research now shows that, given basic personal information, AI can be more persuasive than a human. In the hands of advertisers, political campaigns or propagandists, that is a powerful tool for changing what people believe and do.
How it happens
- Chatbots that use your age, politics and interests to tailor their arguments.
- Millions of personalised conversations running at once — impossible for humans to match.
- Engagement-optimised systems designed to keep you talking, buying or clicking.
Real cases
A 2025 study in Nature Human Behaviour matched about 900 people in debates with either a human or GPT-4. When the AI had basic personal information about its opponent, it was more persuasive than humans 64% of the time; without that information, it performed about the same as people.
Source: Phys.org ↗
Warning signs
- A conversation that seems to know exactly which buttons to press.
- Feeling pushed toward a purchase, vote or belief by a “friendly” chatbot.
What you can do
- Be aware that chatbots can be designed to persuade, not just inform.
- Share as little personal information as you can with AI systems.
- Support transparency rules for AI used in advertising and politics.
Frequently asked questions
What is the danger of manipulation in AI?
AI can hold a conversation, learn who you are and shape its arguments to fit you personally. Research now shows that, given basic personal information, AI can be more persuasive than a human. In the hands of advertisers, political campaigns or propagandists, that is a powerful tool for changing what people believe and do.
How does it happen?
Chatbots that use your age, politics and interests to tailor their arguments. Millions of personalised conversations running at once — impossible for humans to match. Engagement-optimised systems designed to keep you talking, buying or clicking.
How can I protect myself?
Be aware that chatbots can be designed to persuade, not just inform. Share as little personal information as you can with AI systems. Support transparency rules for AI used in advertising and politics.
Latest updates on manipulation
No updates in this category yet — check back tomorrow.
What do you think?
Seen this danger up close? Worried about it? Share it with the community.
Share your thoughts on manipulation