The Hindu Editorial Analysis- 31st March 2023 | Current Affairs & Hindu Analysis: Daily, Weekly & Monthly - UPSC PDF Download

Why in News?

OpenAI has announced GPT4, the latest version of its large language model that powers key applications like ChatGPT and the new Bing.

What is ChatGPT?

• ChatGPT is an artificial-intelligence chatbot developed by San Francisco-based AI research company OpenAI in 2022.
• It is a trained model which interacts in a conversational way. The dialogue format makes it possible to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.
• It can have conversations on topics from history to philosophy, generate lyrics and suggest edits to computer programming code.

Technology Used

• The technology that underlies ChatGPT is referenced in the second half of its name, GPT, which stands for Generative Pre-trained Transformer. 
• Transformers are specialized algorithms for finding long-range patterns in sequences of data. 
• A transformer learns to predict not just the next word in a sentence but also the next sentence in a paragraph and the next paragraph in an essay. This is what allows it to stay on topic for long stretches of text.

Shortcomings of the Previous Models

• The chatbot isn’t always accurate: Their sources aren’t fact-checked, and they rely on human feedback to improve its accuracy. They can also get facts mixed up and produce misinformation.
• GPT-3 and ChatGPT’s GPT-3.5 were limited to textual input and output, meaning they could only read and write. 
• GPT-3 and GPT-3.5 only operated in one modality, text, meaning users could only ask questions by typing them out.

What is GPT-4?

• GPT-4 is a large multimodal model which means it can encompass more than just text and accepts images as inputs.
• GPT-4 also exhibits human-level performance on various professional and academic benchmarks.
• For example, it can answer tax-related questions, schedule a meeting among three busy people, or learn a user’s creative writing style.
• GPT-4 is also capable of handling over 25,000 words of text, opening up a greater number of use cases that now also include long-form content creation, document search and analysis, and extended conversations.

How is GPT-4 different from GPT-3?

• GPT-4 can see images now: GPT-4 can be fed images and asked to output information accordingly whereas GPT-3 and ChatGPT’s GPT-3.5 were limited to textual input and output.
• GPT-4 can process a lot more information at a time.
• GPT-4 has an improved accuracy: GPT-4 significantly reduces hallucinations relative to previous models and scores 40 per cent higher than GPT-3.5 on factuality evaluations. 
• Hard to trick: It will be a lot harder to trick GPT-4 into producing undesirable outputs such as hate speech and misinformation.
• Better at understanding languages other than English: GPT-4 is more multilingual which means that users will be able to use chatbots based on GPT-4 to produce outputs with greater clarity and higher accuracy in their native languages.