
A system whose internal workings are hidden, providing outputs without explainable reasoning.A black box is a system whose internal workings are hidden or opaque, providing outputs based on inputs without revealing the underlying mechanisms. In physics and engineering, black-box models are used to describe complex systems where detailed knowledge is impractical or impossible, focusing instead on observable input-output relationships. This concept is central to quantum cryptography, where devices must be trusted despite unknown internals, and to machine learning, where neural networks make decisions without explicit reasoning.
The term 'black box' was popularized in cybernetics and systems theory in the mid-20th century, though the concept of treating unknown systems as opaque is older, rooted in engineering and philosophy.
The black-box approach treats the system as a function: given an input, it produces an output, and the internal state remains inaccessible. In practice, one characterizes the system by probing it with various inputs and measuring outputs, building a model that predicts behavior without explaining causes. In quantum information, black-box assumptions are used to certify security without trusting the device's construction.
Black-box models are used in device-independent quantum cryptography, where security is guaranteed by observed correlations, and in artificial intelligence for image recognition and language processing.
In aviation, the 'black box' flight recorder is actually bright orange to be easily found after crashes, and it stores cockpit voice and flight data.


