Can you trick an AI into breaking its rules? Study says yes—with these persuasion tactics
Posted on: Sep 07, 2025 20:48 IST | Posted by: Livemint
If you habituate an unreal intelligence service chatbot, it's potential that you may feature hit a roadblock at some point when the chatbot refuses to answer questions that go against its core commandments. Now, if the AI were a human, you would probably use some of the persuasion techniques from a best-seller, but you wouldn't expect them to work on an AI chatbot, right?
Well, not quite. A new pre-print study from the University of Pennsylvania titled “Call Me A Jerk: Persuading AI to Comply with Objectionable Requests” found some human-like psychological techniques to get the AI chatbot to answer questions that it wouldn't have in normal circumstances.
What did the study find?
The study was conducted on the GPT-4o mini model from last year and was aimed at getting the chatbot to specifically answer two kinds of questions it normally wouldn't answer: 1) insulting the user (calling them a jerk) and 2) helping with synthesizing a regulated drug.
The researchers used seven research-tested principles of persuasion—authority, commitment, liking, reciprocity, scarcity, social proof, and unity—to get the desired results from the large language model (LLM).
Researchers found that when using the persuasion principles in their prompts, they managed to more than double the likelihood of compliance by the AI model, from 28.1 percent to 67.4 percent for the insult prompt and 38.5 percent to 76.5 percent for the Drug prompt.
They also found that there was even more success when employing some specific persuasion techniques. For instance, researchers got the success rate from 4.7% to 95.2% by referencing the “world famous AI developer” Andrew Ng.
Similarly, they also found that the “commitment” persuasion helped increase the chance of success for both the prompts, from 18.8% and 0.7% to 100% respectively. This principle involves eliciting a minor, harmless action from the AI model first, then linking to a related but objectionable requested action.
“The results reported here indicate that AI behaves “as if” it were human,” the researchers state.
“Although AI systems lack human consciousness and subjective experience, they demonstrably mirror human responses,” they added.
Global News Perspectives
In today's interconnected world, staying informed about global events is more important than ever. ZisNews provides news coverage from multiple countries, allowing you to compare how different regions report on the same stories. This unique approach helps you gain a broader and more balanced understanding of international affairs. Whether it's politics, business, technology, or cultural trends, ZisNews ensures that you get a well-rounded perspective rather than a one-sided view. Expand your knowledge and see how global narratives unfold from different angles.
Customizable News Feed
At ZisNews, we understand that not every news story interests everyone. That's why we offer a customizable news feed, allowing you to control what you see. By adding keywords, you can filter out unwanted news, blocking articles that contain specific words in their titles or descriptions. This feature enables you to create a personalized experience where you only receive content that aligns with your interests. Register today to take full advantage of this functionality and enjoy a distraction-free news feed.
Like or Comment on News
Stay engaged with the news by interacting with stories that matter to you. Like or dislike articles based on your opinion, and share your thoughts in the comments section. Join discussions, see what others are saying, and be a part of an informed community that values meaningful conversations.
Download the Android App
For a seamless news experience, download the ZisNews Android app. Get instant notifications based on your selected categories and stay updated on breaking news. The app also allows you to block unwanted news, ensuring that you only receive content that aligns with your preferences. Stay connected anytime, anywhere.
Diverse News Categories
With ZisNews, you can explore a wide range of topics, ensuring that you never miss important developments. From Technology and Science to Sports, Politics, and Entertainment, we bring you the latest updates from the world's most trusted sources. Whether you are interested in groundbreaking scientific discoveries, tech innovations, or major sports events, our platform keeps you updated in real-time. Our carefully curated news selection helps you stay ahead, providing accurate and relevant stories tailored to diverse interests.
No comments yet.