Margaret Rouse is an award-winning technical writer and teacher known for her ability to explain complex technical subjects simply to a non-technical, business audience. Over…
Black box AI is any type of artificial intelligence (AI) that is so complex that its decision-making process cannot be explained in a way that can be easily understood by humans. Black box AI is the opposite of explainable AI (XAI).
Causes of black box AI include:
Proprietary IT – the inner workings of an AI model are kept secret to protect intellectual property.
Deep learning – Deep neural networks (DNNs) and deep learning algorithms create thousands (and sometimes millions) of non-linear relationships between inputs and outputs. The complexity of the relationships makes it difficult for a human to explain which features or interactions led to a specific output.
Black box AI is undesirable for a number of reasons. When the internal workings of an AI system are not understood, it becomes increasingly challenging to identify why an AI model is producing biased outputs and where errors in logic are occurring. It also makes it difficult to determine who should be held accountable when outputs are flawed or outright dangerous.
When an AI system is transparent and interpretable, it becomes easier to trust in the integrity of the system and the accuracy of its outputs. Transparency and interpretability can be achieved through a variety of approaches, including designing and using algorithms that are easily understood by humans, making sure that human feedback always plays a role in the decision-making process and developing tools that are able to provide visual explanations for how an AI application arrives at a decision.
Popular tools being developed to prevent black box AI and ensure responsible AI include:
Editor’s Note: According to the Apple Style Guide, writers should avoid the term black box and use the terms closed box or opaque box instead. The writer for this definition would like to propose “mystery box” as a substitute label.
Techopedia’s editorial policy is centered on delivering thoroughly researched, accurate, and unbiased content. We uphold strict sourcing standards, and each page undergoes diligent review by our team of top technology experts and seasoned editors. This process ensures the integrity, relevance, and value of our content for our readers.
Margaret is an award-winning technical writer and teacher known for her ability to explain complex technical subjects to a non-technical business audience. Over the past twenty years, her IT definitions have been published by Que in an encyclopedia of technology terms and cited in articles by the New York Times, Time Magazine, USA Today, ZDNet, PC Magazine, and Discovery Magazine. She joined Techopedia in 2011. Margaret's idea of a fun day is helping IT and business professionals learn to speak each other’s highly specialized languages.
What is Turnitin AI Checker? The Turnitin AI checker is an advanced tool aimed at maintaining the integrity of school...
Maria WebbTechnology journalist
What is ISO/IEC 42001? ISO/IEC 42001 is an international standard that provides a governance framework for implementing and continually improving...
Margaret RouseTechnology Expert
What are Physical Resource Networks (PRNs)? The definition of Physical Resource Networks (PRNs) is that they are a type of...
Nicole WillingTechnology Journalist
Trending NewsLatest GuidesReviewsTerm of the Day