Explained: What is ChatGPT, an AI chatbot which can give 'every answer'

ChatGPT, a new chatbot developed by artificial intelligence non-profit OpenAI Inc., is the talk of the town as users wonder about its intelligence to provide a detailed answer on historical arguments to poems on cryptocurrency.

What is ChatGPT?

It is a trained chatbot which interacts conversationally. The dialogue format, OpenAI claims, makes it possible for ChatGPT to answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.

It is said that it can respond to everything - from stories, mathematical solutions to theoretical essays.

Background of OpenAI and its Elon Musk connection

OpenAI is a California-based company famous for its GPT-3 software, which releases AI models that reply to text prompts. The company’s recently released second version of its DALL-E model went viral for its skill to bring about photo-realistic images just from user prompts.

OpenAI was co-founded by Tesla Inc. CEO Elon Musk and Sam Altman - the present CEO, with other investors, about seven years ago. But in 2018, because of differences, Musk left the company. However, Twitter's new boss was found endorsing the model’s abilities on the micro- logging website.

ALSO READ: Now, you can write articles using artificial intelligence tool, like this

Is Chatbot a novel technology?

Chatbot technology is certainly not new. Several such models, including some from tech behemoths like Meta and Microsoft, have been released but with mixed success.

Microsoft’s AI chatter bot named ‘Tay’ - ‘Thinking About You’- was released on Twitter in March, 2016, under the name TayTweets and handle @TayandYou. But it was soon taken down after it posted racist, sexist and offensive comments learnt from Twitter users.

Meta’s Blender bot 3 released this year took it to the next level. It reportedly replied to a prompt about the company’s CEO Mark Zuckerberg that his company exploits people for money and he doesn’t care. It needs to stop!

Bugs reported in OpenAI's ChatGPT

Users point the model is giving a detailed but incorrect answer to a question on algebra, and its intelligence to override limits on results related to issues like gore, crime and racism.

(Inputs from Bloomberg)