ChatGPT is reportedly scraping Google Search data to answer your questions - here's how
Briefly

OpenAI leverages Google Search index data to assist ChatGPT in answering user queries, with emphasis on current events. The data use concentrates on news, sports, and financial markets where in-house models can produce inaccurate responses. OpenAI obtains search result data through a third-party web-scraping provider, SerpApi, which markets extraction services for building machine learning datasets. A former Google engineer ran an experiment by creating a fabricated search term, indexing it on a hidden page in Google, and observing ChatGPT surface related information, indicating access to Google-indexed content. The approach affects search traffic and content attribution dynamics.
As more people consult ChatGPT for general inquiries, reports are pointing to OpenAI, the AI-powered chatbot's parent company, leveraging Google Search's index to provide users with answers. OpenAI is using Google Search data to help ChatGPT answer questions, particularly about news, sports, and financial markets, according to . The article said that these current event topics are where OpenAI's proprietary tools struggle to give accurate responses.
The Information also reported that OpenAI retrieves Search data from SerpApi, a web-scraping firm. According to 's website, the company offers data extraction services to AI models, saying that search results data best builds large datasets for machine learning models. Abishek Iyer, a former Google engineer, conducted an experiment to prove ChatGPT uses Google Search indexes to generate information. In his experiment, , Iyer created a fake word, included it on a hidden page,
Read at ZDNET
[
|
]