Tricks for making AI chatbots break rules are freely available online – New Scientist

Advertisement
Explore by section
Explore by subject
Explore our products and services
Certain prompts can encourage chatbots such as ChatGPT to ignore the rules that prevent illicit use, and they have been widely shared on social platforms
By Chris Stokel-Walker
21 August 2023

AI chatbots have rules written into their code to prevent illicit use – but they can be subverted
Jamie Jin/Shutterstock
Prompts that can encourage chatbots like ChatGPT to ignore pre-coded rules have been shared online for more than 100 days without being patched, potentially enabling people to use the bots for criminal activity.
Artificial intelligence-based chatbots are given a set of rules by their developers to prevent misuse of the tools, such as being asked to write scam emails for hackers. However, because of the conversational nature of the technology, it is possible …
Advertisement
To continue reading, subscribe today with our introductory offers
No commitment, cancel anytime*
Offer ends 14th June 2023.
*Cancel anytime within 14 days of payment to receive a refund on unserved issues.
Inclusive of applicable taxes (VAT)
Existing subscribers
Advertisement
Explore the latest news, articles and features
Features
Subscriber-only
Analysis
Subscriber-only
News
Subscriber-only
Insight
Subscriber-only
Trending New Scientist articles
1
2
3
4
5
6
7
8
9
10
Advertisement
Download the app

source