OpenAI, the company that developed ChatGPT chatbot, got an early lead in generative artificial intelligence (AI) technology. However, a report has now claimed that its probable toughest rival Google is pretty much well equipped to compete with it, especially after it upgraded its Bard chatbot with a new machine-learning model post I\/O 2023.
Google announced PaLM 2, its state-of-the-art language model with improved multilingual, reasoning and coding capabilities in May this year.The company has since been making regular updates in problem solving capabilities of its AI chatbot Bard.
According to a report by the Information, while Google develops a sequel to PaLM 2, \u201cit may hold a trump card: YouTube.\u201d The Google-owned streaming website is arguably \u201cthe single biggest and richest source of imagery, audio and text transcripts on the internet.\u201d Citing a person with knowledge of the situation, the report said that Google\u2019s researchers have been using YouTube to develop its next large-language model, Gemini. It is to be noted that Google CEO Sundar Pichai talked about how Google brought together DeepMind and Google Brain to make Google DeepMind, which is using computational resources and building more capable systems, safely and responsibly.
\u201cGemini was created from the ground up to be multimodal, highly efficient at tool and API integrations and built to enable future innovations, like memory and planning,\u201d he said, noting that it offers \u201cimpressive multimodal capabilities not seen in prior models.\u201d
OpenAI reportedly trained its AI models on YouTube<\/span>
YouTube as a data source is a hot property, and the report also claimed that OpenAI secretly used data from the platform to train some of its AI models, one person with direct knowledge of the effort was quoted as saying.<\/p>\nIt is to be noted that YouTube\u2019s terms of service forbid using content for anything other than \u201cpersonal, non-commercial use.\u201d However, it is known that in the AI industry everyone is scraping the web constantly.
The development comes about two and half months after media reports said that Google used ChatGPT data to train Bard. It was also claimed that Google used data from OpenAI\u2019s ChatGPT, scraped from a website called ShareGPT, to train Bard.
The Pixel smartphone maker, however, denied saying \u201cBard is not trained on any data from ShareGPT or ChatGPT.\u201d<\/p>\n<\/div>\n