How we tricked AI chatbots into creating misinformation, despite ‘safety’ measures

By Lin Tian, Research Fellow, Data Science Institute, University of Technology Sydney
Marian-Andrei Rizoiu, Associate Professor in Behavioral Data Science, University of Technology Sydney

When you ask ChatGPT or other AI assistants to help create misinformation, they typically refuse, with responses like “I cannot assist with creating false information.” But our tests show these safety measures are surprisingly shallow – often just a few words deep – making them alarmingly easy to circumvent.

We have been investigating how AI language models can be manipulated to generate coordinated disinformation campaigns across social media platforms. What we found should concern anyone worried about the integrity of online information.

The shallow safety problem

Read complete article

~	Your say: week beginning September 1

~	Local journalists and fixers are dying at unprecedented rates in Gaza. Can anyone protect them?

~	New data shows a jump in older people dying from drug overdoses

~	Australian parents are helping their kids buy a first home with less money, but more rent-free living

~	Over-the-top melodrama, a platonic rom-com and retirement village murders: what to watch in September

~	From daycare through to uni, we’ve all had this type of teacher

~	The Pacific’s united front on climate action is splintering over deep-sea mining

~	How can the International Criminal Court achieve justice for women?

~	NZ’s shift to more private healthcare will likely raise costs and reduce quality: what the evidence tells us

~	Report on social media age assurance trial says there is not a one-size-fits-all solution