AI Glossary for Journalists

Also known as: AI fluency, AI competency

The ability to understand, critically evaluate, and effectively use artificial intelligence tools and concepts. For journalists, AI literacy encompasses knowing how AI works at a conceptual level, being able to use AI tools productively, understanding AI's limitations and biases, and making informed decisions about when and how to apply AI in journalistic work.

Also known as: Explainability, AI openness

The practice of being open and clear about how AI systems work, what data they use, how decisions are made, and when AI is involved in content creation. For journalism, AI transparency means disclosing AI use to audiences, documenting AI-assisted processes, and advocating for transparency from AI companies and institutions that deploy AI systems affecting the public.

Also known as: Computational procedure

A set of step-by-step instructions or rules that a computer follows to solve a problem or complete a task. In AI, algorithms are the mathematical procedures that allow machines to learn from data and make predictions or decisions.

Also known as: Algorithmic accountability reporting

The principle that organizations using algorithms and AI systems should be held responsible for the outcomes those systems produce. In journalism, this has a dual meaning: newsrooms must be accountable for their own AI use, and journalists should investigate how algorithms used by governments, corporations, and platforms affect the public.

Also known as: Application Programming Interface

Application Programming Interface — a set of protocols that allows different software applications to communicate with each other. In journalism, APIs enable newsrooms to integrate AI capabilities directly into their workflows, content management systems, and publishing tools. For example, using OpenAI's API to build custom AI tools tailored to newsroom-specific tasks.

Also known as: AI

The broad field of computer science focused on creating systems capable of performing tasks that typically require human intelligence, such as understanding language, recognizing patterns, making decisions, and generating content. Modern AI in journalism primarily refers to machine learning systems and large language models.

Also known as: Algorithmic journalism, robot writing

The use of algorithms and AI to automatically generate news content from structured data without direct human writing. Common applications include financial earnings reports, sports game recaps, weather updates, and election results. The Associated Press pioneered this approach in 2014 for corporate earnings stories.

Also known as: Algorithmic bias, AI bias

Systematic errors or skewed perspectives in AI systems that result from biased training data, flawed assumptions in algorithm design, or unrepresentative data collection. In journalism, AI bias can lead to stereotypical portrayals, underrepresentation of certain communities, or skewed analysis that reflects historical inequities rather than current reality.

Also known as: Conversational AI, AI assistant

A software application designed to simulate conversation with human users, typically through text-based interfaces. Modern AI chatbots like ChatGPT, Claude, and Gemini use large language models to generate contextually relevant responses and can assist journalists with research, writing, and analysis tasks.

Also known as: Computer-assisted reporting (CAR)

The application of computer science methods — including AI, data analysis, and automation — to journalism practice. It encompasses data journalism, automated reporting, algorithmic accountability reporting, and the use of computational tools for investigation and storytelling. It represents the broader integration of computing into journalistic workflows.

Also known as: Data-driven journalism, DDJ

A practice of journalism that uses data collection, analysis, and visualization to uncover and tell stories. AI enhances data journalism by enabling reporters to analyze larger datasets, identify patterns more quickly, and create more sophisticated visualizations — even without advanced statistical or coding skills.

Also known as: Deep neural networks

A subset of machine learning that uses neural networks with many layers (hence 'deep') to learn complex patterns from large amounts of data. Deep learning powers most modern AI applications in journalism, including language models, image recognition, and speech-to-text systems.

Also known as: Synthetic video, face-swap video

Synthetic media — typically video or audio — created using deep learning AI to realistically depict people saying or doing things they never actually said or did. Deepfakes pose significant challenges for journalism: they can be used to spread misinformation, manipulate public opinion, and undermine trust in authentic video evidence.

Also known as: Verification

The process of verifying the accuracy of claims, statements, and information before or after publication. In the AI era, fact-checking takes on added importance as AI tools can generate plausible-sounding but false information. AI can assist the fact-checking process by helping organize verification tasks, but cannot replace human verification against primary sources.

Also known as: Model fine-tuning, domain adaptation

The process of taking a pre-trained AI model and further training it on a specific, smaller dataset to specialize its performance for particular tasks. A newsroom could fine-tune a language model on its own article archive to better match its style guide, terminology, and editorial standards — creating a customized AI assistant.

Also known as: GenAI, generative models

AI systems that can create new content — text, images, audio, video, or code — based on patterns learned from training data. For journalists, generative AI tools like ChatGPT and Claude can draft text, while tools like DALL-E and Midjourney generate images. The 'generative' aspect distinguishes these from AI that only analyzes or classifies existing content.

Also known as: Confabulation, AI fabrication

When an AI system generates information that sounds plausible and authoritative but is factually incorrect, fabricated, or nonsensical. AI hallucinations are particularly dangerous in journalism because they can include fake quotes, non-existent sources, fabricated statistics, and invented events — all presented with confidence.

Also known as: Model inference, prediction

The process of using a trained AI model to generate outputs — predictions, text, images, or decisions — from new inputs. When you type a prompt into ChatGPT and receive a response, the model is performing inference. Understanding inference matters for journalists because it affects response speed, cost, and the privacy implications of where your data is processed.

Also known as: LLM

A type of AI model trained on vast amounts of text data that can understand, generate, and manipulate human language. LLMs power tools like ChatGPT (GPT-4), Claude (Anthropic), and Gemini (Google). They work by predicting the most likely next word in a sequence, which allows them to generate coherent, contextually relevant text.

Also known as: ML

A branch of artificial intelligence where computer systems learn to perform tasks by analyzing patterns in data rather than being explicitly programmed with rules. In journalism, machine learning is used for content recommendation, audience analytics, automated content classification, and pattern detection in large datasets.

Also known as: NLP

The branch of AI focused on enabling computers to understand, interpret, and generate human language. NLP powers many journalism-relevant tools including search engines, translation services, sentiment analysis, text summarization, and chatbots. It's the technology that allows you to interact with AI using everyday language.

Also known as: Artificial neural network, ANN

A computing system inspired by the biological neural networks in the human brain, consisting of interconnected nodes (neurons) organized in layers that process information. Neural networks learn by adjusting the strength of connections between nodes based on training data. They form the backbone of modern AI systems used in journalism.

Also known as: AI prompt, input query

The text input or instruction given to an AI system to generate a response. In journalism, effective prompts include specific instructions, context about the task, the desired output format, and the role the AI should play. The quality of the prompt directly affects the quality of the AI's output — a concept known as 'prompt engineering.'

Also known as: Retrieval-Augmented Generation

Retrieval-Augmented Generation — a technique that enhances AI responses by first retrieving relevant information from a specific knowledge base or document collection, then using that information to generate more accurate and contextual answers. For newsrooms, RAG can connect AI tools to internal archives, style guides, or verified databases, reducing hallucinations and improving accuracy.

Also known as: Ethical AI, trustworthy AI

An approach to developing and deploying AI systems that prioritizes ethical considerations, fairness, transparency, accountability, and the minimization of harm. For newsrooms, responsible AI means using AI tools in ways that uphold journalistic standards, protect sources, avoid bias, maintain accuracy, and serve the public interest.

Also known as: Bot journalism, automated reporting

A colloquial term for automated journalism — the use of AI algorithms to generate news articles without direct human authorship. Despite the name, no physical robots are involved; the term refers to software that converts structured data into narrative text. It is most commonly used for high-volume, formulaic content such as financial reports and sports scores.

Also known as: AI-generated media

Any media content — including text, images, audio, and video — that is generated or significantly modified by AI. This encompasses AI-written articles, generated images, cloned voices, and deepfake videos. For journalists, understanding synthetic media is crucial both for creating content ethically and for detecting manipulated media in the information landscape.

Also known as: Text tokenization

The process of breaking text into smaller units called tokens — which can be words, parts of words, or characters — that AI models can process. Understanding tokenization helps journalists know why AI tools have input limits, why costs vary by text length, and why some languages require more tokens (and thus more processing) than others.

Also known as: Training set, training corpus

The collection of text, images, or other information used to teach an AI model. For large language models, training data typically includes books, websites, articles, and other text sources. The quality, diversity, and biases present in training data directly shape the AI's capabilities and limitations — including any biases it may exhibit.

Also known as: Transformer architecture

A neural network architecture introduced in 2017 that revolutionized AI by enabling models to process entire sequences of text simultaneously and understand relationships between words regardless of their distance in a sentence. Transformers are the foundation of all modern large language models — the 'T' in GPT stands for Transformer.

Also known as: Vector store, embedding database

A specialized database that stores data as mathematical vectors (numerical representations), enabling efficient similarity searches. In journalism, vector databases power AI features like semantic search across article archives, finding related stories, and enabling RAG systems that connect AI tools to newsroom-specific knowledge bases.