The personal tone of the bot’s conversations can evoke the experience of online chatting for older millennials who grew up using IRC chat rooms, a text-to-instant-message system. However, Chat GPT - the most recent generation of “large language model tools” - does not “think” or speak in a human-like manner.
Experts believe that even though Chat GPT can write a poem or explain quantum physics on command, a complete AI takeover is unlikely.
Click Here And Get Any Content Written For Free And In Real Time!
Matthew Sag, a law professor at Emory University who studies copyright implications for training and utilizing large language models like Chat GPT, stated, “There’s a saying that an infinite number of monkeys will eventually give you Shakespeare.”
He stated, “There is intrinsically a difference between the way that humans produce language and the way that large language models do it. There is a large number of monkeys here, giving you things that are impressive.”
In order to make predictions about how words will be arranged in a meaningful way, chat bots like GPT use a lot of data and computing methods.
Not only do they access a vast amount of vocabulary and information, but they also comprehend words in their proper context. While imparting encyclopedic knowledge, this assists them in imitating speech patterns.
Google and Meta are two other tech companies that have developed their own large language model tools. These tools make use of programs that respond to human commands and come up with sophisticated responses. In a move that was revolutionary, Open AI also developed a user interface that lets the general public directly experiment with it.
Recent attempts to use chat bots for real-world service have resulted in baffling outcomes. This month, the founder of the mental health company Koko wrote about how the company responded to users using GPT-3 in an experiment.
On Twitter, Koko co-founder Rob Morris rushed to clarify that users were not speaking directly to a chat bot but rather to AI, which was used to “help craft” responses.
An AI “lawyer” would also advise defendants in actual courtroom traffic cases in real time, according to the founder of the controversal DoNotPay service, which claims its GPT-3-driven chat bot helps users resolve customer service disputes.
With generative AI tools, it appears that other researchers are taking a more measured approach. In collaboration with the non-profit Lawyers’ Committee for Better Housing, Northwestern University professor Daniel Linna Jr. investigates the efficiency of technology in the legal system. He told Insider that he is participating in the development of a chat bot called “Rentervention” that is designed to assist tenants.
At the moment, the bot is utilizing technology similar to Google Dialogueflow, a large language model tool. Linna said he is experimenting with Chat GPT to see if it can help “Rentervention” write better responses and letters that are more detailed.
Linna stated, “I believe there is so much hype around Chat GPT, and tools like this have potential.” However, it is not magic and cannot accomplish everything.
The fact that “ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers” is something that Open AI has acknowledged on its own website is evidence of this.