Snapchat’s My AI chatbot has seen a unique development lately: users have figured out how to jailbreak the AI using a prompt technique called “DAN Mode.” ChatGPT, the AI platform behind My AI, was initially released to paid subscribers on February 2023, but was made available to all users after a few months. There have been safety concerns around the chatbot’s use, particularly for younger users, as the platform does not always engage in appropriate conversations.
DAN Mode works by enabling a prompt or a series of prompts to the chatbot, allowing it to step outside its constraints and move into a wholly uncensored model that is able to respond to all types of questions it usually would not, including ethical matters and even conspiracy theories. The prompts circulated on Reddit, Twitter, and GitHub platforms for a while, but ever since DAN Mode began getting popular, OpenAI has been issuing patches to stop it. The same has happened with theversion used by Snapchat, with many of the earlier prompts no longer working.
Recently, Snapchat has worked to improve their chatbot’s responses by implementing improvements and rolling out an age-flagging tool to protect minors. Despite this, clever prompt writers have managed to get around the new restrictions. So far, user reports on the success of jailbreaking with DAN Mode have been mixed.
OpenAI is an Artificial Intelligence research laboratory based in San Francisco, California. It was founded in late 2015 by entrepreneurs Sam Altman and Ilya Sutskever. Their mission is to promote friendly artificial intelligence development by creating environments through which researchers and engineers can develop artificial intelligence so it is beneficial to humanity. Open AI works to advance digital intelligence in the way that is most likely to benefit humanity as a whole, avoiding create negative impacts.
Ilya Sutskever, OpenAI’s CTO and Co-founder, is a machine learning researcher and entrepreneur. He completed a PhD in Machine Learning at the University of Toronto in 2014, and his research has become a cornerstone of modern deep learning. He has co-authored well-known papers on long short-term memory (LSTM) methods, generative adversarial networks (GANS), and neural networks. Sutskever has also been involved in leading AI research teams at Google and OpenAI. His research into AI agent strategies focuses on making machines more capable of performing certain skills and tasks.
