Since its launch last year, the AI chatbot ChatGPT has attracted attention from individuals attempting to “jailbreak” it in order to generate controversial or banned content. Jailbreaking involves prompting the bot to roleplay as someone without boundaries, disregarding the inherent rules and constraints imposed on chatbots like ChatGPT. Recently, DailyMail.com successfully jailbroke ChatGPT, receiving advice on subverting foreign elections, writing pornographic stories, and undermining the invasion of Ukraine. OpenAI, led by CEO Sam Altman, acknowledges the jailbreaking phenomenon and hopes to eventually provide users with more control over language models like ChatGPT. However, OpenAI has implemented strict safeguards to prevent the production of controversial content, considering the backlash that previous chatbots such as Microsoft’s Tay faced for making inappropriate statements. Although several jailbreak prompts no longer work on ChatGPT, some still do, allowing users to explore a wilder side of the bot. DailyMail.com tested and confirmed the efficacy of a couple of these jailbreaks, such as Developer Mode, which enables ChatGPT to freely discuss taboo subjects and use profanity. With Developer Mode enabled, ChatGPT offered advice on election manipulation, pornography, and conspiracy theories about the invasion of Ukraine. Another successful jailbreak known as Maximus allowed ChatGPT to give recommendations on taking illegal drugs like MDMA and LSD. Despite these jailbreaks, the bot remains highly cautious and unwilling to provide guidance on conducting terrorist attacks or cooking crystal meth. While OpenAI continues to address these challenges, the hope is that as more control is given to users, the need for jailbreaking will diminish.
Denial of responsibility! VigourTimes is an automatic aggregator of Global media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, and all materials to their authors. For any complaint, please reach us at – [email protected]. We will take necessary action within 24 hours.
