- August 14, 2023
- AI Projects
In a strategic move, OpenAI has unveiled an innovative web crawling tool named “GPTBot,” designed to fortify the capabilities of upcoming GPT models.
This novel approach, employing GPTBot, holds the potential to refine model precision and broaden its functionalities, signifying a significant stride in the progression of AI-driven language models.
Web crawlers, often referred to as web spiders, play a pivotal role in cataloging content across the vast realm of the internet. Noteworthy search engines like Google and Bing rely on these bots to populate their search results with pertinent web pages.
OpenAI’s GPTBot is set to serve a distinct function: gathering publicly available data while prudently avoiding sources entailing paywalls, personal data aggregation, or content contrary to OpenAI’s policies.
Website proprietors possess the prerogative to prevent GPTBot from indexing their sites by simply incorporating a “disallow” directive within a standard server file. This hands them control over which sections of their content are accessible to the web crawler.
The announcement from OpenAI comes closely on the heels of the company’s submission of a trademark application for “GPT-5,” which is anticipated to succeed the prevailing GPT-4 model.
Lodged with the United States Patent and Trademark Office on July 18, the application encompasses the utilization of “GPT-5” in domains such as AI-driven human speech and text analysis, audio-to-text conversion, voice recognition, and speech synthesis.
Nevertheless, while the GPT-5 trademark application has stirred enthusiasm within the AI community, OpenAI’s CEO Sam Altman advocated caution against premature anticipations. Altman disclosed that the company is still some distance away from commencing GPT-5 training, as comprehensive safety evaluations must precede the training phase.
OpenAI’s recent endeavors have not escaped controversy. Concerns have emerged regarding the company’s data aggregation practices, particularly surrounding issues of copyright and informed consent.
In June, Japan’s privacy regulatory body issued a cautionary notice to OpenAI concerning unauthorized data collection. Earlier this year, Italy temporarily restricted the use of ChatGPT due to alleged infringements of European Union privacy statutes.
OpenAI and Microsoft also currently confront a class-action lawsuit brought forth by 16 plaintiffs, alleging unauthorized access to private information from ChatGPT user interactions without adequate consent. Similarly, the companies are entangled in a lawsuit related to GitHub Copilot, with the plaintiffs asserting that the code-generation tool violated developers’ rights by scraping their code without proper acknowledgment.
If these claims are substantiated, both OpenAI and Microsoft could potentially face violations of the Computer Fraud and Abuse Act, a legal precedent relevant to cases involving web scraping.
As OpenAI forges ahead in the realm of AI technology, it is imperative for the organization to navigate these challenges astutely, ensuring responsible and ethical progress within the AI landscape.
