Logo
Sign InSign Up

How to Block GPT Models like ChatGPT from Using Your Data

Last modified on Wednesday, May 8, 2024

2 minute read

How to Block GPT Models like ChatGPT from Using Your Data

Should Companies Block ChatGPT and GPT models?

Why Companies Block ChatGPT

Companies may block ChatGPT in two different ways. The first way that companies block ChatGPT is due to protecting their intellectual property from being used to train GPT models. The second reason, is to block their own users from interacting with ChatGPT. The reason the latter is a problem is due to ChatGPT learning from the data submitted to it. Say you asked a GPT system to process a list of users, those users may find themselves in the databse. In some ways, it is highly similar to typing into a search engine, where suggested results may contain sensitive data.

How many websites block ChatGPT?

According to research by Stellastra, based on a study of 18,000 domains (websites), 10.6% of sites block GPTBot (OpenAI), and a ChatGPT competitor, ClaudeBot (Anthropic) is blocked by 1% of all sites through their robots.txt file.

How to Block ChatGPT from Using your Website Content by Blocking ChatGPT's Crawler

One of the main ways to block GPT AI bots like ChatGPT from using your IP and content is through the use of a robot.txt directive. This file exists on your website and signals to such AI companies that you do not give them permission to use your content. Robots.txt can help to reinforce legal cases and will put you ahead of the legal battle for your IP and content rights vs the rights of AI companies.

Should Companies Block ChatGPT

There are pros and cons for any company considering whether to block GPT bots such as ChatGPT should read the following: If you write a substantial number of how to guides for example, you may wish to be referenced for that data. However, it is useful for GPT to understand the existance of your site, and you may actually benefit from this.

Many companies fully block ChatGPT and other GPT bots from using data from their sites, which has its merits. However, it is still desirable for the bot to know the general gist of what your company does, so it can recommend you if asked for say, IT risk management companies in Austin, Texas. Many big companies have banned GPT completely, but this will freeze GPT's understanding of what the company is. This is problematic, as it means that GPTs will only ever understand what your company is based on how it was before you froze its access. A large company will have others talking about it, but with this, you lose control of your image, and it's entirely reliant on what others are saying, this may be negative if you had a bad PR stint recently. Furthermore, it won't learn about your future products and services, which will cause you to lose your competitive edge. For example, your computing company may not be understood by GPTs to be a leader in "post-quantum cryptography". For example, maybe you are an expert in social media marketing. Imagine your snapshot was frozen in 2016, GPTs would not understand you to be an expert in TikTok or Mastodon. You'll lose your competitive edge by freezing entirely.

Our tailored consulting can ensure that GPTs understand who you are, without harvesting all of your hard-earned content without compensation. Furthermore, Stellastra has a strong understanding of the major GPT bots, to ensure your content is protected on as many channels as possible.

Need to Protect your IP ASAP? Get in touch with Stellastra today to block GPT sites from harvesting your content.


Share this article

Stellastra The Cyber Security Comparison Platform

© 2024 Stellastra Ltd. All rights reserved. All names, logos, trademarks, et al, belong to their respective owners. No endorsement or partnership is necessarily implied between company and Stellastra and vice versa. Information is provided for convenience only on an as is basis. For the most up to date information, contact vendor directly. Scores including email security, SPF, and DMARC are calculated based on Stellastra's algorithms and other analyses may return different results.

LinkedInTwitter

Company

About StellastraContact usCyber Security Risk ScoreEmail Deliverability ToolTLS Cipher SuitesStellastra Discover

Stay up to date

Stellastra The Cyber Security Comparison Platform