Stories
OpenAI not to use customer data to train AI? How will this effect chatGPT-4 and chatGPT-5?
OpenAI changes its training plan.
OpenAI changes its training plan.
In a statement, the CEO of OpenAI said, “OpenAI has changed its plans as the customers don’t want AI to get trained on their data.” On Friday, CEO Sam Altman mentioned OpenAI has not trained its AI large-language models as GPT by paying for customer data for a while.
The terms of service of OpenAI were updated on March 1. ChatGPT was trained earlier on customer data and was used to train its large-language models and ChatGPT. The plans have changed as the customers don’t want AI to get trained on their data. Microsoft, Salesforce and Snapchat are OpenAI’s business customers. These are more likely to take advantage of OpenAI’s API capabilities.
AI stands for Artificial Intelligence. In 2015, OpenAI was founded, also known as American Artificial Intelligence. Ilya Sutskever, Vicki Cheung, Greg Brockman, Trevor Blackwell, Durk Kingma, John Schulman, Andrej Karpathy, Pamela Vagata, and Wojciech Zaremba, with Elon Musk and Sam Altman serving as the initial board members found OpenAI.
What is ChatGPT?
ChatGPT stands for Chat Generative Pre-trained Transformer. ChatGPT follows the instructions in a prompt that provides a detailed response. It is a sibling model or a general-purpose trained chatbot. ChatGPT uses artificial intelligence to generate text after a user enters a prompt, developed by tech startup OpenAI.
The chatbot uses GPT-4, a large language model which uses deep learning to produce human-like text. ChatGPT also automates chat tasks and is considered an intelligent and helpful chatbot.
How are chatbot and chatGPT different?
ChatGPT generates responses based on the context and tone of the conversation, whereas, Chatbots are pre-programmed with a limited set of responses.
According to the reports, Amazon recently warned its employees and asked them not to share confidential information with ChatGPT fearing it might show up to everyone in the answers as the AI trains on customer data which further led to the update in terms and service of AI, introducing new privacy and data protection which is only visible to customers or users who use the company’s API services.
Training on customer data happens when the users or customers of AI enter any information on the wildly popular chatbot ChatGPT.
What is API?
The word Application in API refers to any software with a distinct function. In Application Programming Interface or API, an interface is a contract of service between two applications.
The communication happens between the two using requests and responses. In simple words, API is a piece of technology that allows two different systems to communicate with each other.
Different types of API
The four commonly used APIs in web service are private, public, partner and composite.
One example of the most widely used API is the Google Maps API.
Barry Diller, an American businessman, Chairman, and Senior Executive of IAC and Expedia Group, has suggested media companies take their issues to the courts and potentially sue AI companies over the use of creative content. Companies can easily take their issues to the courts with the executives being equally concerned about the impact of ChatGPT and other AI platforms on their intellectual property.
About GPT 5 and its updates
According to reports, OpenAI is developing GPT-5. Released this March, GPT-5 is the presumed successor to its AI language model GPT-4. Although GPT-5 is currently not trained, it will be expanding the capabilities of GPT-4
Considering the safety implications of such work, Altman said that the team is doing other things on top of GPT-4. The team will work and address all sorts of safety issues.
It will be interesting seeing how AI is upgrading itself for the future. People will be looking for Altman’s further statement on OpenAI, GPT-5 and the challenges and security of AI.
Published by: Aliya Khan