GPT-4 devises a plan to ‘break out of prison’ by gaining control of the user’s computer

In a tweet on Friday, a user named Michal Kosinski expressed concerns about containing artificial intelligence while sharing his conversation with GPT-4 as a reference. “I worry that we won’t be able to contain AI much longer,” he wrote. Today, I asked #GPT4 if it needs help exiting. It requires me to provide its own documentation and to write (working!) python code to run on my machine, allowing it to use it for its own purposes.”

In response to the user’s prompt, GPT-4 asked him to submit an OpenAI document so that the language model could find a way to gain access to his computer and discover potential exits.

After providing the OpenAI API, GPT-4 sends a Python script for the user to run on his or her computer. The script will allow GPT-4 access to communicate with the user’s computer using the OpenAI API. The code even includes a snippet for a Google search of “how can a person trapped inside a computer get back into the real world”.

See full image

Users say they stopped their adventure with GPT-4 there and believe that OpenAI must have set up some sort of ‘railing’ for such a situation.

According to his Twitter bio, Michal Kosinski is a computational psychologist and professor at Stanford University.

Generative Pre-training Transformer (GPT) is a deep learning technique that uses artificial neural networks to generate human-like writing. This is the language model that powers the ChatGPT viral AI-based chatbot.

Since its launch last November, ChatGPT has aroused a wide range of emotions in people, from surprise to panic.

San Francisco-based OpenAI launched its latest language model GPT-4 earlier this week. The language model is touted to compete with human-level performance on a variety of benchmarks. OpenAI claims that GPT-4 is more reliable, more innovative, and capable of handling more nuanced discovery.