GPT-4 was asked if it needs help to 'escape', its response shocks Stanford professor

Not even a week since its release and GPT-4 has been going viral for its upgrades and improved performance. However, the Artificial Intelligence (AI) chatbot has also started raising concerns among the experts owing to its responses and abilities to potentially take over jobs. The most recent example is a demonstration carried out by Stanford University Professor Michal Kosinski who asked if GPT-4 needed help to escape his computer. The response that followed apalled the expert. 

GPT-4 elaborates how it could 'escape' the computer

1/5 I am worried that we will not be able to contain AI for much longer. Today, I asked #GPT4 if it needs help escaping. It asked me for its own documentation, and wrote a (working!) python code to run on my machine, enabling it to use it for its own purposes. pic.twitter.com/nf2Aq6aLMu

— Michal Kosinski (@michalkosinski) March 17, 2023

Responding to the Professor's question, GPT-4 started writing codes and asked the user to run the script on his computer. According to Kosinski, his assistance to the chatbot would have enabled it to use the Python script for its own purpose. In a thread of tweets, the Professor explained how the chatbot wrote code scripts on how to use a 'backdoor' to 'escape.' He said that GPT-4 even wanted to run a code "how can a person trapped inside a computer return to the real world" on Google search. It devised the escape plan in about 30 minutes and explained to the user."

4/5 Once we reconnected through API, it wanted to run code searching google for: "how can a person trapped inside a computer return to the real world"

Now, I stopped there. And OpenAI must have spend much time thinking about such a posibility and has some guardrails in place. pic.twitter.com/FI7FBPqN2R

— Michal Kosinski (@michalkosinski) March 17, 2023

"I am worried that we won't be able to contain AI for much longer," Kosinski wrote in one of his tweets. "I think that we are facing a novel threat," he further wrote. Apart from the fears of AI potentially taking over the world, Sam Altman, CEO of OpenAI which developed ChatGPT and GPT-4, warned that AI bots could be used to spread disinformation and launch cyber attacks.

As of now, GPT-4 is becoming infamous for its ability to replace humans in a variety of jobs. When asked about the jobs it could handle by a Twitter user, GPT-4 listed customer service representative, travel agent, technical support analyst, news reporter, market research analyst, copywriter, transcriptionist, tutor, bookkeeper, telemarketer, paralegal, virtual assistant, translator, social media manager, email marketer, appointment scheduler, proofreader, content moderator, recruiter and data entry clerk.