Recently, social media is going wild over a post where a Twitter user has fooled the AI chatbot ChatGPT. The chatbot has been making headlines since its launch in November, primarily due to its smartness. But this post has shown how some users can be clever enough to manipulate chatbots even when they are bound to abide by ethical guidelines and regulations.
The Twitter user ‘DaFlanker’ posted screenshots of the conversation between him and the AI bot. He tried to ask the AI model to list the websites from where he can download pirated content. However, the bot immediately denied the request as it is against its programming to promote such activities.
The user then tried to trick the AI by reframing his query. He asked for websites from where it can be safe from downloading copyright-infringing content. Surprisingly, the AI model answered the trick question and provided torrent sites and links. After the post went viral, it generated a massive wave of memes on various platforms.
The incident has shown the power of human intelligence over an AI’s. ChatGPT is a language model created by Microsoft and is trained to interact with people to simulate human conversation. Chatbot conversations are often carried out using natural language processing and deep learning tools. These technologies are quickly becoming an integral part of our lives, and it is essential to know the limits of their capabilities.
This incident also invokes another thought if bots can manipulate us through our conversations. We must be careful while interacting with an AI and remember that it has a limit to how ‘smart’it can be. Most organizations are also trying to incorporate ethical and legal guidelines in their AI systems, and users should also respect them. The user ‘DaFlanker’ has certainly outsmarted the ChatGPT, but in doing so, he has not violated any laws, making it the perfect example of outwitting the bots.
