Last Updated on
ChatGPT, the AI language model from OpenAI, brings artificial intelligence to database management systems. However, ChatGPT’s query capabilities may be limited by the link, or lack thereof, between it and an SQL server instance. Let’s learn how the OpenAI LLM (Large Language Model) can perform SQL queries with natural language inputs. Can ChatGPT query a database?
Can ChatGPT access a database?
No, ChatGPT cannot access a database – not directly, at least.
ChatGPT Plus and ChatGPT Enterprise, the paid subscription tiers for ChatGPT, allow the user to install plugins. Through plugins, the AI language model can interface with other software’s, and perform tasks on behalf of the humans that ordinarily use it.
Essential AI Tools
Jasper AI
Copy.ai
Winston AI detector
Originality AI detector
WordAI
The ChatGPT plugin selection includes those for data management and data analytics purposes. ChatGPT Excel plugins exist for this very purpose.
Can you connect ChatGPT to a database?
ChatGPT cannot connect to database by itself. However, with ChatGPT Plus or ChatGPT Enterprise, the AI chatbot can access the internet and any publicly available databases on it. Google Sheets can be used to create publically accessible databases for an AI chatbot, which includes Microsoft Bing Chat and Google Bard.
Can ChatGPT query a database?
Yes, ChatGPT can query a database, by using the ChatGPT API (or OpenAPI as its lesser known). The OpenAI API allows development of custom apps and functionality, powered by the GPT-3.5 Turbo or GPT-4 model.
SQL developers who know a little Python can run real-time analytics and complex queries using human-like text with prompt engineering. Prompt engineering is the art of knowing how to ask for what you want.
How to use ChatGPT to query a database
Data analysts will have a head start here, being that the first step in using ChatGPT to query a database is to know how to query any database.
Thankfully, AI takes a lot of the technical know-how out of it. You don’t need to be a software engineer to use AI for database management – in truth, you barely need to know SQL syntax or other query languages, but your knowing your database schema does help.
The way ChatGPT works is by taking an input prompt in natural human language, and feeding this into its NLP (Natural Language Processing) algorithm. It can then produce a text output – which again, can be in natural human language, or it can be code. SQL (Structured Query Language) is one such example. This behaviour can be used for various applications, including generating a MySQL or PostgreSQL script.
This isn’t an invitation to get lazy, however, as a broad AI client like ChatGPT can make mistakes. It’s best to use ChatGPT for guidance, while still using SQL server management studio in conjunction.