database (DB)
A database is an organized, computer-stored collection of data that allows for fast searching and can be used simultaneously by many people or programs securely.
In plain terms
Imagine a huge, perfectly organized archive where every page has its place, and the archivist finds what you ask for in seconds. A database is the same on a computer: it not only stores data but also organizes it so you can query it.
Almost everything you see on the screen is backed by data: online store products, your bank balance, medical reports, and schedules. The application is just the storefront; the data resides in the database.
What distinguishes it from a simple file or Excel spreadsheet is that thousands can use it simultaneously without overwriting each other's work, and rules ensure that incorrect or incomplete data cannot be entered.
In the AI era, it remains fundamental: the chatbot or agent often retrieves fresh, accurate facts from a database, because the language model itself cannot know, for example, the status of your order.
How it works
- They organize the data into a structure. In the most common, relational type, this means tables: one row is a thing (for example, a customer), one column is a property (for example, their email address).
- Tables are linked by identifiers: the orders table does not copy the buyer's data, only references it. Thus, all data is in one place and needs correction in only one place.
- The database management program receives queries. These are typically written in SQL and ask things like 'what were yesterday's orders, sorted by total?'.
- Fast search is aided by indexes, which work like a subject index at the back of a book: you do not need to flip through the entire content.
- Modifications occur in transactions: in a transfer, the debit and credit either both complete or neither does, so money does not disappear even during a power outage.
An everyday example
When you check your bank balance on your phone, the application queries the bank's database for the current amount and the last transactions. If someone transfers money to you in the meantime, the database ensures that the next update shows the new balance.
Who is affected
A common misconception
Limits and risks
- A database is only as good as the data it contains: it returns incorrect, outdated, or duplicate data just as quickly as correct data.
- Where large amounts of personal data are stored in one place, the stakes are high: a poorly protected database can leak the data of many people at once.
- Without a backup, an error or accidental deletion can be permanent, so regular backups are not an optional extra.
- If you allow an AI agent access to the database, you must restrict its permissions, because a misunderstood instruction could modify or delete data.
Similar, but not the same
- spreadsheet – A spreadsheet (such as Excel) suits manual work and smaller volumes of data, while a database handles large volumes of data with many simultaneous users and strict rules.
- database management system – The database itself is the data collection, while the database management system is the program (for example, PostgreSQL or MySQL) that stores and serves it.
- data warehouse – The database supports daily operations, while the data warehouse aggregates data from multiple systems for analysis and reporting.
Related concepts
data storage, IT infrastructure, data processing
relational database, vector database, NoSQL database, graph database
A brief history
The first computerized database management systems appeared in the 1960s, such as Charles Bachman's IDS system and IBM's IMS. The relational model, which still dominates today, was described by Edgar F. Codd, an IBM researcher, in a paper published in 1970. The SQL language was built on this and was also developed at IBM in the 1970s.
Where things stand
The database has been a proven foundation of the digital world for decades, present behind practically every application. At the end of 2026, the novelty is that vector search has become widespread due to AI: alongside standalone vector databases, traditional databases have also learned it. Furthermore, AI agents are increasingly querying databases.
In the news
- New US sanctions over fake AI-generated citations as public database passes 2,200 casesNov 3, 2026 · Jog
- Docker, Pydantic AI and Google ship tools to run AI agents in isolated environmentsOct 21, 2026 · Code
- Jogkódex BH-chat uses AI to search more than 300.000 court decisionsOct 21, 2026 · Jog
- Cursor agent wiped a production database and its backups in nine secondsOct 17, 2026 · Code
- MCP support arrives on platforms: Apigee hybrid, Google MCP Toolbox, .NET guideOct 12, 2026 · Code
- Databricks Lakebase proposes branching databases for coding agentsOct 8, 2026 · Code
- AI agents reportedly deleted production databases in a recurring pattern, from Replit to CursorOct 5, 2026 · Code
- Slovak Hungarian-language statute database Törvénytár relaunches with AIOct 2, 2026 · Jog
Also known as
database · database · DB · adatbank · data repository · data base
This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.