


OpenAI releases ChatGPT plug-in, causing concern: content quality and security need to be measured
News on March 24, on Thursday, local time in the United States, the artificial intelligence research company OpenAI released multiple plug-ins for its popular chat robot ChatGPT, which will help ChatGPT connect to the network to Retrieve information in real time or use third-party applications.
It is reported that OpenAI’s plug-in will allow ChatGPT to access third-party knowledge sources and databases, including the Internet, thus greatly expanding the capabilities of the chatbot. OpenAI will provide the Alpha version to ChatGPT test application users and developers. Priority will initially be given to a small number of developers and ChatGPT Plus subscribers before being made accessible to the general public.
The most attractive plug-in is undoubtedly OpenAI’s first-party web browsing plug-in, which allows ChatGPT to find data from the web to answer various questions posed to it. Previously, ChatGPT’s information content was limited to dates, events and people before September 2021. The new plug-in uses Microsoft's Bing search API to retrieve content from the web, while displaying the websites it visited when summarizing the answer and citing its source in ChatGPT's response.
However, being able to access other information sources online may present new risks, as previous research by the OpenAI team has found. OpenAI created an experimental system called WebGPT in 2021, which sometimes cited information from unreliable sources and was induced to pick data from websites that users may trust, even if these sources are not objectively authoritative. Meta's later offline BlenderBot 3.0 also had access to the Internet, but it quickly deviated from the original vision. When prompted by certain texts, it posts conspiracy theories or other offensive content.
Live networks are less accurate than static training data sets, and this means there are fewer filtering restrictions. Search engines like Google and Bing use their own security mechanisms to reduce the chance of unreliable content rising to the top of search results, but these results can manipulated, they are not necessarily representative of the entire network. As previously reported, Google’s algorithm prioritizes sites that use modern web technologies such as encryption, mobile support, and schema markup. Therefore, many websites with other high-quality content tend to be missed. This gives search engines a lot of power to cherry-pick data that might inform answers to web language models. Google was found to be prioritizing its own services in search, such as answering travel queries with data from Google Places, rather than aggregating information from richer, more social sources like TripAdvisor. At the same time, algorithmic approaches to search are also prone to loopholes. In 2020, Pinterest took advantage of an anomaly in Google's image search algorithm to show more content in Google Image Search. OpenAI acknowledged that an internet-enabled ChatGPT could be used for all types of bad behavior, such as sending fraudulent messages and spam emails, bypassing security restrictions, and generally "enhancing the capabilities of those who deceive, mislead, or abuse others." capabilities of bad actors”. But the company also said it had "implemented several safeguards" internally and externally to prevent this from happening.
In addition to the browser plug-in, OpenAI also provides the model behind ChatGPT with a Python interpreter that works in a sandbox, firewall execution environment, and some temporary disk space. This plugin supports uploading files to ChatGPT and downloading the results. OpenAI says it's particularly useful for solving mathematical problems, performing data analysis, visualization, and converting between file formats.
Many early collaborators have built plugins for ChatGPT, including Expedia, FiscalNote, Instacart, Kayak, Klarna, Milo, OpenTable, Shopify, Slack, Speak, Wolfram, and Zapier.
For the most part, the purpose of these plugins is self-explanatory. For example, the OpenTable add-on will allow the chatbot to search restaurants for available reservations, while the Instacart add-on allows ChatGPT to place orders from local stores. Zapier is one of the most scalable apps ever created, connecting with apps like Google Sheets, Trello, and Gmail to trigger a range of productivity-related tasks.
To encourage developers to create new plug-ins, OpenAI has open sourced a "retrieval" plug-in that enables ChatGPT to access document snippets from data sources such as files, notes, emails, or public documents by asking questions in natural language. . OpenAI posted: "We are working hard to develop various plug-ins and bring them to a wider audience. We still have a lot to learn, and with your help, we hope to create something that is both practical and safe."
For ChatGPT, plug-ins are an essential supplement. ChatGPT was once strictly limited to learning from its training data, but with the plug-in, it suddenly became more powerful and perhaps less legally risky. Many experts have accused OpenAI of profiting from training ChatGPT on unlicensed data. ChatGPT's dataset contains a wide variety of public websites, and the plugin may solve this problem by allowing companies to retain full control over their data. (Xiao Xiao)
The above is the detailed content of OpenAI releases ChatGPT plug-in, causing concern: content quality and security need to be measured. For more information, please follow other related articles on the PHP Chinese website!

Hot AI Tools

Undresser.AI Undress
AI-powered app for creating realistic nude photos

AI Clothes Remover
Online AI tool for removing clothes from photos.

Undress AI Tool
Undress images for free

Clothoff.io
AI clothes remover

Video Face Swap
Swap faces in any video effortlessly with our completely free AI face swap tool!

Hot Article

Hot Tools

Notepad++7.3.1
Easy-to-use and free code editor

SublimeText3 Chinese version
Chinese version, very easy to use

Zend Studio 13.0.1
Powerful PHP integrated development environment

Dreamweaver CS6
Visual web development tools

SublimeText3 Mac version
God-level code editing software (SublimeText3)

Hot Topics



How to define header files using Visual Studio Code? Create a header file and declare symbols in the header file using the .h or .hpp suffix name (such as classes, functions, variables) Compile the program using the #include directive to include the header file in the source file. The header file will be included and the declared symbols are available.

Writing C in VS Code is not only feasible, but also efficient and elegant. The key is to install the excellent C/C extension, which provides functions such as code completion, syntax highlighting, and debugging. VS Code's debugging capabilities help you quickly locate bugs, while printf output is an old-fashioned but effective debugging method. In addition, when dynamic memory allocation, the return value should be checked and memory freed to prevent memory leaks, and debugging these issues is convenient in VS Code. Although VS Code cannot directly help with performance optimization, it provides a good development environment for easy analysis of code performance. Good programming habits, readability and maintainability are also crucial. Anyway, VS Code is

Running Kotlin in VS Code requires the following environment configuration: Java Development Kit (JDK) and Kotlin compiler Kotlin-related plugins (such as Kotlin Language and Kotlin Extension for VS Code) create Kotlin files and run code for testing to ensure successful environment configuration

Depending on the specific needs and project size, choose the most suitable IDE: large projects (especially C#, C) and complex debugging: Visual Studio, which provides powerful debugging capabilities and perfect support for large projects. Small projects, rapid prototyping, low configuration machines: VS Code, lightweight, fast startup speed, low resource utilization, and extremely high scalability. Ultimately, by trying and experiencing VS Code and Visual Studio, you can find the best solution for you. You can even consider using both for the best results.

VS Code is absolutely competent for Java development, and its powerful expansion ecosystem provides comprehensive Java development capabilities, including code completion, debugging, version control and building tool integration. In addition, VS Code's lightweight, flexibility and cross-platformity make it better than bloated IDEs. After installing JDK and configuring JAVA_HOME, you can experience VS Code's Java development capabilities by installing "Java Extension Pack" and other extensions, including intelligent code completion, powerful debugging functions, construction tool support, etc. Despite possible compatibility issues or complex project configuration challenges, these issues can be addressed by reading extended documents or searching for solutions online, making the most of VS Code’s

Sublime Text is a powerful customizable text editor with advantages and disadvantages. 1. Its powerful scalability allows users to customize editors through plug-ins, such as adding syntax highlighting and Git support; 2. Multiple selection and simultaneous editing functions improve efficiency, such as batch renaming variables; 3. The "Goto Anything" function can quickly jump to a specified line number, file or symbol; but it lacks built-in debugging functions and needs to be implemented by plug-ins, and plug-in management requires caution. Ultimately, the effectiveness of Sublime Text depends on the user's ability to effectively configure and manage it.

Of course! VS Code integrates IntelliSense, debugger and other functions through the "C/C" extension, so that it has the ability to compile and debug C. You also need to configure a compiler (such as g or clang) and a debugger (in launch.json) to write, run, and debug C code like you would with other IDEs.

VS Code's task system improves development efficiency by automating repetitive tasks, including build, test, and deployment. Task definitions are in the tasks.json file, allowing users to define custom scripts and commands, and can be executed in the terminal without leaving VS Code. Advantages include automation, integration, scalability, and debug friendliness, while disadvantages include learning curves and dependencies. Frequently asked questions include path issues and environment variable configuration.
