Can you install ChatGPT without worrying? Are ChatGPT plugins safe and secure?
In this ChatGPT plugin tutorial, we are going to cover if ChatGPT plugins are safe to use, what are the reasons that prove that ChatGPT plugins are safe and what measurements and actions OpenAI has taken to make ChatGPT plugins safe and secure.
Are ChatGPT Plugins Safe To Install?
Yes, based on our experience ChatGPT plugins are safe to install.
OpenAI – the company behind ChatGPT – has prioritized safety and security when they have developed the ChatGPT plugin features as you can read the details below.
We also researched Reddit, online forums and online publications and we haven’t seen any safety concerns regarding ChatGPT plugin users or reporters covering ChatGPT.
Another reason that you can trust ChatGPT plugins is that the majority of the ChatGPT plugins currently available in the ChatGPT plugin store is developed by well-known trusted companies like Expedia, Zapier, etc.
If you are still unsure if it is safe to install ChatGPT plugins, make sure to install ChatGPT plugins of the company you are familiar with.
You can check the complete list of ChatGPT plugins here.
OpenAI Safety Measures Regarding ChatGPT Plugins
Safety is of paramount importance when it comes to utilizing ChatGPT plugins.
OpenAI acknowledges that connecting language models to external tools introduces both new opportunities and significant risks.
Plugins offer the potential to address challenges associated with large language models, such as staying up-to-date, accessing proprietary information sources, and strengthening responses with evidence-based references.
ChatGPT plugin safety measures and safeguards
To ensure the safety and reliability of ChatGPT plugins, OpenAI has implemented several measures and safeguards. From the inception of the plugin platform, OpenAI has focused on mitigating risks and has performed extensive red-teaming exercises both internally and with external collaborators.
These exercises have uncovered possible scenarios where plugins, if released without safeguards, could perform prompt injection, send fraudulent emails, bypass safety restrictions, or misuse information sent to the plugin.
OpenAI utilizes these findings to inform safety-by-design mitigations that restrict risky plugin behaviors and enhance transparency in the user experience.
The goal is to ensure that plugins operate securely, providing users with reliable and trustworthy results while minimizing the risk of unintended consequences.
ChatGPT Plugins gradual deployment as it related to safety
OpenAI has also made the decision to gradually deploy access to plugins, enabling them to carefully evaluate and address any potential safety challenges that may arise.
OpenAI actively encourages researchers interested in studying safety risks and mitigations in this area to participate in their Researcher Access Program. This collaborative effort allows for a thorough exploration of the safety implications of ChatGPT plugins and aids in developing effective mitigation strategies.
Additionally, developers and researchers are invited to submit plugin-related safety and capability evaluations using OpenAI’s open-sourced Evals framework.
Is ChatGPT Browsing Plugin safe to use?
In terms of specific plugins, OpenAI has developed the browsing plugin as an experimental model that can access and retrieve information from the internet.
This plugin is designed with safety as a top priority, operating as a good citizen of the web. It utilizes a text-based web browser that is limited to making GET requests, reducing certain classes of safety risks.
Browsing activities are conducted through the Bing search API, leveraging Microsoft’s work on source reliability, truthfulness of information, and “safe-mode” to prevent the retrieval of problematic content.
OpenAI respects content creators and adheres to web norms by configuring the browser plugin to honor websites’ robots.txt files, providing transparency in the browsing process.
