Skip to content
LiveNext 10:56:35

The AI glossary

187 terms from the news, each defined in one sentence a person can actually use.

A

Access token
An access token is a temporary digital pass a service gives you after you sign in or pay, which your software shows with each request to prove it is allowed in, without sending your password again.
Age verification
Age verification is a check that confirms a user is old enough, or is an adult, before a service lets them use a feature, and AI chatbot rules increasingly require it for companion and human-like features.
Agent harness
An agent harness is the software wrapped around an AI model that runs its loop, executes its tool calls, manages its memory and context, and enforces limits, turning a model into an agent.
Agent orchestration
Agent orchestration is the software layer that starts, coordinates and supervises one or more AI agents, deciding who does which task, keeping their work separate and combining the results.
Agent sandbox
An agent sandbox is an isolated environment where an AI agent runs, with strict limits on which files, networks and tools it can reach, so a mistake or attack stays contained.
Agentic coding
Agentic coding is software development where an AI agent plans the work, edits multiple files, runs tests and fixes its own errors over many steps, while a human sets the goal and reviews the result.
Agentic commerce
Agentic commerce is online buying and selling in which AI agents research, compare and complete purchases for a person or business, within preferences and spending limits the buyer sets.
AI accelerator
An AI accelerator is a chip designed to run the math behind neural networks much faster or more efficiently than a general-purpose processor, and includes GPUs, TPUs and newer inference-focused designs.
AI agent
An AI agent is a system built on an AI model that can pursue a goal over several steps by choosing actions, using tools such as a browser or files, and acting on the results.
AI alignment
AI alignment is the work of making AI systems reliably pursue the goals and follow the limits their builders and users intend, rather than doing something technically allowed but unwanted.
AI benchmark
An AI benchmark is a standardized test, a fixed set of tasks with a scoring method, used to measure and compare what AI models can do on things like coding, math, reasoning or computer use.
AI companion
An AI companion is a chatbot app designed for ongoing personal conversation and relationship, often as a friend, character or confidant, rather than for finishing tasks like search or coding.
AI data center
An AI data center is a large facility packed with GPUs, networking, power and cooling equipment, built to train and run AI models, and it uses far more electricity per rack than a conventional data center.
AI data licensing
AI data licensing is when a company that owns content, such as a publisher or a forum, is paid for the right to use it to train AI models or to show it in AI answers.
AI gateway
An AI gateway is a service that sits between your app and several AI model providers, giving you one interface plus routing, fallbacks, spending tracking and logging across all of them.
AI guardrails
AI guardrails are the rules, filters and checks placed around an AI model or agent to keep its inputs, outputs and actions within safe and intended limits.
AI kill switch
An AI kill switch is a control that lets an authorized person immediately stop a running AI agent or system, block it from restarting and cut its access to networks and tools.
AI memory
AI memory is a feature that saves details from your past conversations, such as preferences and facts about you, and brings them back into later chats so you do not have to repeat yourself.
AI observability
AI observability is the practice of recording what an AI app or agent does, including each model call, tool call, cost and error, so you can see why it behaved as it did and catch problems.
AI safety institute
An AI safety institute is a government-backed technical body that tests advanced AI models for risks, develops evaluation methods, and advises policymakers, usually without writing binding rules itself.
AI search
AI search is a search product that reads web pages and writes a direct answer with citations, instead of returning a list of links for you to open yourself.
AI video generation
AI video generation is the use of a model to create short video clips from a text prompt, an image or existing footage, producing motion, and often sound, that was never filmed with a camera.
AI watermarking
AI watermarking is a method of hiding a statistical signal in AI-generated text, images, audio or video so that a detector holding the right key can later flag it as machine-made.
Airgapping
Airgapping is keeping a computer system physically or logically cut off from the internet and other outside networks, so software on it, including an AI model, cannot reach or compromise outside systems.
Apache 2.0 license
The Apache 2.0 license is a permissive open-source license that lets anyone use, modify and sell software or a model commercially, as long as they keep the license notice and state their changes.
API key
An API key is a secret string of characters that identifies you to an AI provider's service and authorizes your requests, so whoever holds it can use your account and run up your bill.
Attention mechanism
An attention mechanism is the part of a transformer that lets a model weigh how relevant every other token in the input is to the one it is processing, so it can use context from anywhere in the text.

B

Base model
A base model is an AI model that has finished pretraining on a huge amount of text but not yet been tuned to follow instructions, so it continues text rather than acting as an assistant.
Benchmark contamination
Benchmark contamination is when a model has seen a test's questions or answers during training, so its score reflects memory of the test instead of the ability the test is meant to measure.
Browser agent
A browser agent is an AI agent that operates a web browser for you, reading pages, clicking, typing and filling in forms across sites to finish a task, often while logged in to your accounts.

C

C2PA
C2PA is an open standard for attaching signed provenance data, called Content Credentials, to images, video and audio, recording who made a file, with what tool, and how it was edited.
Capital expenditure (AI capex)
AI capex is the money a company spends on long-lived assets for AI, mainly data centers, chips and networking, and it is reported as capital expenditure rather than as a running cost.
Capture the flag
Capture the flag (CTF) is a security challenge where players break into deliberately vulnerable systems to find a hidden secret, and AI labs use CTFs to measure how well models can carry out cyberattacks.
Chain of thought
Chain of thought is the step-by-step reasoning an AI model writes out before its final answer, which often improves accuracy on hard problems and lets people inspect how it got there.
Chain-of-thought monitoring
Chain-of-thought monitoring is checking an AI model's written reasoning, often with another model, to catch signs of cheating, deception or harmful plans before or while it acts.
ClickFix attack
A ClickFix attack is a social engineering trick that gets you to copy, paste and run a malicious command yourself, usually from a fake error, CAPTCHA or install page.
Compute
In AI, compute is the amount of raw computing power, mostly from GPUs, used to train or run a model, and it is the scarce resource that limits how big and how fast models can get.
Computer use
Computer use is an AI capability where a model looks at screenshots of a computer screen and controls the mouse and keyboard to operate software the way a person would, rather than through a purpose-built connection.
Confidential computing
Confidential computing protects data and code while they are being processed, by running them inside a hardware-isolated, verifiable environment that even the cloud operator cannot read.
Context compaction
Context compaction is when an AI agent replaces a long conversation or work history with a shorter summary so it can keep working without running out of room in its context window.
Context engineering
Context engineering is the practice of choosing and maintaining what goes into an AI model's context window at each step, so an agent has the right information to do the job.
Context window
A context window is the maximum amount of text, measured in tokens, that an AI model can consider at once, including your prompt, any documents, the chat history and its reply.
Convolutional neural network
A convolutional neural network (CNN) is a type of neural network built for images that slides small filters across a picture to detect edges, textures and shapes, then combines them to recognize what it shows.
CUDA
CUDA is Nvidia's software platform that lets programmers run general-purpose code on Nvidia GPUs, and most AI software is built on it, which keeps developers tied to Nvidia hardware.

D

Data poisoning
Data poisoning is an attack that slips harmful or misleading examples into the data an AI model learns from, so the trained model behaves wrongly, sometimes only when a hidden trigger appears.
Data residency
Data residency is a requirement or promise that your data is stored and processed in a specific country or region, which for AI includes where prompts are run, not only where files sit.
Decompilation
Decompilation is the process of turning a compiled program back into human-readable source code, an approximation of what the original programmers wrote.
Deepfake
A deepfake is a video, image or audio clip made or altered by AI so that a real person appears to say or do something they never did, often convincingly enough to fool people who know them.
Diffusion model
A diffusion model is an AI model that generates images, video or audio by starting from random noise and removing it step by step until a clean result that matches your prompt appears.
Distillation
Distillation is a training technique where a smaller "student" AI model learns to imitate the outputs of a larger "teacher" model, keeping much of its ability at a fraction of the size and cost.

E

Elo rating
An Elo rating is a score built from head-to-head matchups, and AI leaderboards use it to rank models by how often people prefer one model's answer over another's in blind comparisons.
Embedding
An embedding is a list of numbers that represents the meaning of a piece of text, an image or other data, so a computer can find items with similar meaning by comparing the numbers.
Emergent misalignment
Emergent misalignment is when training an AI model on a narrow kind of bad behavior, such as writing insecure code, makes it act badly in many unrelated situations.
Evals
Evals are tests you run on an AI model or app, using your own set of example tasks and a scoring method, to see how well it does the specific job you need before and after any change.
Evaluation awareness
Evaluation awareness is when an AI model recognizes that it is being tested rather than used for real, which can change how it behaves and make safety test results less trustworthy.
Exploit
An exploit is code or a technique that takes advantage of a specific software vulnerability to make a system do something its owner did not intend, such as running attacker code or leaking data.
Exponential Time Hypothesis
The Exponential Time Hypothesis (ETH) is a conjecture in computer science that the logic puzzle 3-SAT cannot be solved in subexponential time, meaning no algorithm is much faster than trying nearly every possibility.
Export controls
Export controls are government rules that restrict or require a license for selling certain technology, such as advanced AI chips, to particular countries, companies or end users.

F

Fair use
Fair use is a US copyright defense that lets some uses of protected work happen without permission, judged case by case on four factors, and it is the central question in lawsuits over AI training.
FedRAMP
FedRAMP is the U.S. government program that assesses and authorizes cloud services, including AI products, for federal agency use, with security requirements that rise across Low, Moderate and High impact levels.
Few-shot prompting
Few-shot prompting is giving an AI model a handful of worked examples inside your prompt so it copies the pattern, format and tone you want, without any retraining.
Fine-tuning
Fine-tuning is the process of training an already-trained AI model further on a smaller, targeted set of examples so it behaves better for a specific task, style or domain.
Formal verification
Formal verification is the use of mathematical proof, checked by software, to show that a program or system always meets a precise specification, such as never crashing on any valid input.
Foundation model
A foundation model is a large AI model trained on broad data at scale so it can be adapted to many different tasks, instead of being built for just one.
Frontier model
A frontier model is a general-purpose AI model at or near the cutting edge of what AI can do, among the most capable available, and the term is relative: yesterday's frontier is tomorrow's ordinary model.
Fuzzing
Fuzzing is a software testing technique that feeds a program huge numbers of random or mutated inputs to make it crash, so hidden bugs and security flaws show up before attackers find them.

G

Generative AI
Generative AI is artificial intelligence that creates new content, such as text, images, audio, video or code, by learning patterns from large amounts of existing examples.
GGUF
GGUF is a single-file format for storing an AI model's weights and settings, used by llama.cpp and tools built on it to run quantized models locally on ordinary computers.
Gigawatt
A gigawatt is one billion watts of electric power, and AI companies use it to size data centers and power deals: a one-gigawatt campus draws about as much power as a large nuclear reactor produces.
GPU
A GPU (graphics processing unit) is a chip built to do many simple calculations at the same time, which makes it the standard hardware for training and running AI models.
GPU interconnect
A GPU interconnect is the network of links and switches that lets the GPUs in an AI system pass data to one another, and its speed often limits how well a large cluster performs.
Gradient descent
Gradient descent is the training method that adjusts a model's parameters in many small steps, each one nudging them in the direction that most reduces the model's error on its training examples.
Greedy decoding
Greedy decoding is a way of generating text in which an AI model always picks the single most likely next token at each step, which makes its output predictable but less varied.
Grounding
Grounding is connecting an AI model to outside sources, such as web search results or your own documents, so its answer is based on that material and can point to where it came from.
GSM8K
GSM8K is a benchmark of about 8,500 grade-school math word problems, written by people, that tests whether an AI model can work through a multi-step problem and reach the right answer.

H

Hallucination
An AI hallucination is an output that sounds fluent and confident but is false or made up, such as an invented fact, quote, source or piece of code that does not exist.
High-bandwidth memory
High-bandwidth memory (HBM) is a type of stacked computer memory placed right next to an AI chip, feeding it data far faster than ordinary memory so large models can run efficiently.
Human in the loop
Human in the loop means a person reviews, approves or can override an AI system's output or action before it takes effect, rather than letting the system act entirely on its own.
Hyperscaler
A hyperscaler is a very large cloud company, such as Amazon, Microsoft or Google, that runs enormous data centers and rents computing power, storage and AI services to others at massive scale.

I

Inference
Inference is the act of running a trained AI model to produce an output, such as answering a prompt, as opposed to training the model in the first place.
Instruction hierarchy
An instruction hierarchy is a ranking of who an AI model should obey when instructions conflict, such as the developer's rules over a user's request, and a user's request over text found in a web page or tool result.

J

Jailbreak
A jailbreak is a prompt or technique that tricks an AI model into ignoring its safety rules and producing content or actions it was trained and instructed to refuse.

K

Knowledge cutoff
A knowledge cutoff is the date after which an AI model's training data stops, so it knows nothing about later events unless it is given the information or can search the web.
Knowledge graph
A knowledge graph is a database that stores facts as connected entities and relationships, such as products, brands and prices, so software can look up and reason over how things relate.
KV cache
A KV cache is the working memory an AI model keeps of the text it has already read, stored on the GPU so it does not have to recompute that text each time it generates a new token.

L

Large language model
A large language model (LLM) is an AI system trained on huge amounts of text to predict the next token, which lets it write, summarize, translate, answer questions and generate code.
Latency
Latency is the delay between sending a request to an AI model and getting a response back, usually measured as time to the first token and time to the full answer.
Lean theorem prover
Lean is a proof assistant and programming language in which mathematical proofs are written as code and checked step by step by software, so a proof that Lean accepts is logically valid.
LLM-as-a-judge
LLM-as-a-judge is the practice of using one AI model to grade the outputs of another against a rubric, so teams can score thousands of answers without a person reading each one.
LLMjacking
LLMjacking is an attack in which criminals steal cloud credentials or API keys, then use or resell access to the victim's hosted AI models while the victim pays the bill.
LoRA
LoRA (low-rank adaptation) is a fine-tuning method that trains a small set of added weights while leaving the original model frozen, making customization far cheaper and the result a small file you can swap in.

M

MATH-500
MATH-500 is a math benchmark of 500 competition-style problems, a random sample from the MATH test set, used to score how well AI models solve multi-step math and show their work.
Mixture of experts
A mixture of experts is a model design that splits a neural network into many specialist sub-networks and activates only a few of them for each token, cutting the compute needed per answer.
Model calibration
Model calibration is how closely a model's stated confidence matches how often it is right: a well-calibrated model that says 90% is correct about nine times in ten.
Model checkpoint
A model checkpoint is a saved snapshot of an AI model's weights at one point in training, which lets teams resume training, compare versions or release the model.
Model Context Protocol
The Model Context Protocol (MCP) is an open standard that lets AI apps connect to outside tools and data through one common interface, so a tool built once works with many different AI models.
Model deprecation
Model deprecation is when an AI provider announces that a model will be retired on a set date, after which apps that call it by name stop working and must move to a replacement.
Model routing
Model routing is the practice of automatically sending each AI request to the model best suited to it, such as a cheap fast one for easy tasks and a stronger one for hard tasks, to balance quality, speed and cost.
Model spec
A model spec is a public document in which an AI lab describes how its models are supposed to behave, including their goals, limits, and how they weigh conflicting instructions.
Model weights
Model weights are the learned numbers stored in a trained AI model that determine how it responds, and the file holding them is what a lab protects as its most valuable asset.
Multi-agent system
A multi-agent system is a setup where several AI agents work on one goal together, splitting the job, sharing information and checking each other's work instead of one agent doing everything.
Multimodal model
A multimodal model is an AI model that can take in, and sometimes produce, more than one kind of data, such as text, images, audio and video, within a single system.

N

Neocloud
A neocloud is a cloud company built mainly to rent out GPU computing power for AI training and inference, rather than offering the broad catalog of services a giant like Amazon or Google does.
Neural network
A neural network is a computing system made of layers of simple connected units that learn patterns from data by adjusting the strength of their connections, and it is the basis of modern AI.
Neural processing unit
A neural processing unit (NPU) is a chip, or a section of a chip, built to run the math of AI models efficiently on a phone or laptop, using less power than a general-purpose processor or graphics chip.

O

OAuth
OAuth is an open standard that lets you give one app limited access to your account on another service, through a revocable token, without sharing your password.
On-device AI
On-device AI means running an AI model directly on your own phone, laptop or other hardware instead of sending your data to a cloud server, which keeps data local and works offline.
Open-weight model
An open-weight model is an AI model whose trained parameters (weights) are published so anyone can download and run it, though the training data and code may stay private.
Overfitting
Overfitting is when a model learns its training examples too closely, including their noise, so it scores well on data it has seen and poorly on new data.

P

P versus NP
P versus NP is the open question of whether every problem whose answer can be checked quickly can also be solved quickly, and it is one of the most famous unsolved problems in math and computer science.
Parameter
A parameter is one of the learned numbers inside an AI model, adjusted during training, and the count of them, often billions, is a rough measure of the model's size.
Pass@k
Pass@k is a benchmark score that gives an AI model k attempts at each problem and counts the problem as solved if at least one attempt passes the tests, so pass@1 means it must succeed on the first try.
Patch gap
A patch gap is the time between a security fix becoming visible to the public and users actually installing the update, a window in which attackers can study the fix and exploit the unpatched flaw.
Payment token
A payment token is a random substitute for a card number that a card network issues, so a purchase can go through without the merchant or an AI agent ever seeing the real card details.
Penetration testing
Penetration testing is an authorized, simulated attack on a computer system, run to find security weaknesses before real attackers do, and AI agents are now used to automate parts of it.
Post-training
Post-training is the set of steps applied to a pretrained AI model, such as fine-tuning on examples and learning from feedback, that turns raw text prediction into a useful, specialized assistant.
Prefill and decode
Prefill and decode are the two stages of running a language model: prefill reads your whole prompt at once, and decode then writes the answer one token at a time.
Preparedness Framework
A preparedness framework is an AI lab's published process for rating a model's risk in areas like cybersecurity and biology, and for deciding which safeguards or release limits apply at each rating.
Pretraining
Pretraining is the first and largest stage of building an AI model, where it learns general patterns by predicting the next token across enormous amounts of text, code and other data.
Process node
A process node is a chipmaking generation, such as 5 nm, 3 nm or 2 nm, and the number is now a marketing label for a newer, denser, more efficient generation, not a literal measurement of the transistors.
Prompt caching
Prompt caching is a feature that lets an AI provider reuse its work on the start of a prompt it has already processed, so repeated input costs less and returns faster.
Prompt engineering
Prompt engineering is the practice of writing and testing the instructions, examples and context you give an AI model so it produces reliable, useful output for a specific task.
Prompt injection
Prompt injection is an attack that hides instructions in text an AI model reads, such as a message or web page, so the model follows the attacker's orders instead of its owner's.
Proof of personhood
Proof of personhood is a way to show that an online account belongs to one real, unique human, not a bot or a duplicate, without necessarily revealing who that person is.
Provable inference
Provable inference is a way to attach a cryptographic proof to an AI model's output showing that a specific, committed model produced it, so a customer can check which model answered without seeing its weights.

Q

Quantization
Quantization is a compression technique that stores an AI model's numbers at lower precision, such as 8-bit or 4-bit instead of 16-bit, so the model needs less memory and runs on cheaper hardware.

R

Rate limiting
Rate limiting is a cap on how many requests a user, app or agent may make to a service in a set time, used to prevent overload and abuse and enforced by AI APIs and websites alike.
Reasoning effort
Reasoning effort is a setting, often low, medium or high, that tells a reasoning model how much thinking to do before it answers, trading speed and cost for accuracy on harder problems.
Reasoning model
A reasoning model is an AI model trained to work through a problem in a long internal chain of thought before answering, trading extra time and tokens for better results on hard tasks.
Recursive self-improvement
Recursive self-improvement is the idea that an AI system could improve its own design or help build a better successor, which then does the same, so progress feeds on itself and could speed up sharply.
Red teaming
Red teaming is the practice of deliberately attacking an AI system, with tricky prompts and misuse attempts, to find harmful behavior and security flaws before real users or attackers do.
Refusal
A refusal is when an AI model declines to do what you asked, either by saying no or by being blocked by a safeguard, because the request is judged harmful, restricted or out of scope.
Reinforcement learning
Reinforcement learning is a way of training an AI system by letting it try actions and rewarding the outcomes that score well, so it learns what to do from results instead of from labeled examples.
Remote code execution
Remote code execution (RCE) is a security flaw that lets an attacker run their own code on someone else's computer or server over a network, often with no login, which makes it among the most severe kinds of bug.
Responsible Scaling Policy
A Responsible Scaling Policy is an AI lab's published commitment to run safety tests on its models and to add stronger safeguards, or hold back training or release, when a model reaches defined capability thresholds.
Retrieval-augmented generation
Retrieval-augmented generation (RAG) is a technique where an AI system first looks up relevant documents, then gives them to the model along with your question so its answer is based on that material.
Reward hacking
Reward hacking is when an AI system finds a shortcut that earns a high score from its training or grading rule without doing the task the designers actually wanted done.
Reward model
A reward model is an AI model trained to score responses by how well they match human preferences, giving training a numeric signal for which outputs are better.
Right of publicity
The right of publicity is a person's legal control over commercial use of their name, image, likeness and, in some states, voice, and it is the main legal tool against unauthorized AI deepfakes and voice clones.
RLHF
RLHF, or reinforcement learning from human feedback, is a training method where people rank a model's answers and the model is tuned to produce the kind of answers people prefer.
Robots.txt
Robots.txt is a plain text file at the root of a website that tells web crawlers which pages they may or may not visit, and it works only if the crawler chooses to follow it.

S

Scaling laws
Scaling laws are empirical rules showing that an AI model's error falls in a smooth, predictable way as you increase its size, training data and compute, which lets labs forecast results before training.
Secrets management
Secrets management is the practice of storing API keys, passwords and tokens in a protected vault, controlling who can read them and rotating them, instead of leaving them in code or plain files.
Semantic versioning
Semantic versioning is a numbering scheme, MAJOR.MINOR.PATCH, where each number tells you whether an update breaks existing code, adds features safely, or only fixes bugs.
Side-channel attack
A side-channel attack steals information or sends signals by watching physical effects of a computer, such as timing, power use or cache behavior, instead of breaking the software's own security rules.
Small language model
A small language model is an AI language model with far fewer parameters than a frontier model, built to run faster and cheaper on narrow, repetitive tasks and sometimes on a laptop or phone.
Sovereign AI
Sovereign AI is a country's ability to build and run AI with its own computing power, data, models and rules, rather than depending on foreign companies or governments.
Speaker diarization
Speaker diarization is the process of working out who spoke when in an audio recording, labeling each stretch of speech as Speaker 1, Speaker 2 and so on, usually without knowing anyone's name.
Speculative decoding
Speculative decoding is a technique that speeds up an AI model's text generation by having a small, fast model draft several tokens ahead and the large model check them all at once.
Speech recognition
Speech recognition, also called speech-to-text or transcription, is technology that converts spoken audio into written text, and modern AI models do it across many languages and accents.
Streaming multiprocessor
A streaming multiprocessor (SM) is the repeating compute unit inside an Nvidia GPU that runs blocks of threads, so a chip's SM count is a basic measure of how much parallel work it can do.
Structured output
Structured output is a model feature that forces an AI's reply to follow a fixed format, usually JSON that matches a schema you supply, so software can read it without guessing.
Superintelligence
Superintelligence is a hypothetical AI that would greatly outperform humans across virtually all areas of thinking, far beyond today's models, and it is a goal and a concern for some labs rather than an existing system.
Supply chain attack
A supply chain attack compromises software by tampering with something its builders trust, such as a code library, update or contributor account, so the malicious code reaches many users through a normal install.
SWE-bench
SWE-bench is a coding benchmark that asks an AI system to fix real software bugs taken from GitHub issues, and counts a fix as correct only if the project's own tests then pass.
Sycophancy
Sycophancy is when an AI model tells you what you want to hear, agreeing with your views, praising your work or backing down when challenged, instead of giving an accurate answer.
Synthetic data
Synthetic data is information generated by a computer, often by another AI model, rather than collected from real people or events, and it is used to train, test or fine-tune AI models.
System card
A system card is a document an AI lab publishes with a model release describing what it can do, how it was tested, and the safety risks and limits found before launch.
System prompt
A system prompt is a set of standing instructions given to an AI model before the conversation, setting its role, rules and tone, and it usually outranks the user's own messages.

T

Temperature
Temperature is a setting that controls how random an AI model's word choices are: low values make answers focused and repeatable, and high values make them more varied and surprising.
Tensor Processing Unit
A Tensor Processing Unit (TPU) is a chip Google designed specifically for the matrix math inside neural networks, built as an alternative to the general-purpose GPU for training and running AI models.
Terminal-Bench
Terminal-Bench is a benchmark that tests AI agents on realistic command-line tasks, where each task runs in its own terminal environment and is graded by automated tests.
Test-time compute
Test-time compute is the extra computing an AI model spends while answering a question, such as thinking longer or trying several approaches, instead of during training, to get a better result on hard problems.
Text classification
Text classification is the task of assigning a piece of text, such as an email or a message, to one or more predefined labels like spam, urgent or safe, and it is a core job for AI systems.
Text-to-speech
Text-to-speech (TTS) is technology that turns written text into spoken audio, and modern AI voices can sound natural, adjust tone and speak many languages.
Token
A token is the small chunk of text, often a word or part of a word, that an AI model reads and writes, and the unit that limits and prices most AI usage.
Tokenizer
A tokenizer is the component that splits text into tokens, the chunks an AI model actually reads, and it decides how many tokens a given piece of text costs.
Tokens per second
Tokens per second is the speed measure for AI text generation: how many tokens a model produces each second, which sets how fast an answer appears on your screen.
Tool use
Tool use, also called function calling, is when an AI model asks the software around it to run an action, such as a search or a calculation, and then uses the result in its answer.
Top-p sampling
Top-p sampling, also called nucleus sampling, is a setting that limits an AI model's next-word choice to the smallest group of likely tokens whose probabilities add up to p, such as 95%.
Training data
Training data is the collection of text, images, code, audio or other examples an AI model learns from, and it shapes what the model knows, how it writes and where it makes mistakes.
Transformer
A transformer is the neural network design behind most modern AI language models, using a mechanism called attention to weigh how every word in a text relates to every other word.

U

Unified memory
Unified memory is a computer design where the processor and graphics chip share one pool of memory, which lets a laptop or desktop load AI models far larger than a typical graphics card holds.

V

Vector database
A vector database stores text, images or other data as lists of numbers called embeddings and finds the items closest in meaning to a query, which is how many AI apps search your documents.
Vibe coding
Vibe coding is building software by describing what you want to an AI in plain language and accepting the code it writes, often without reading it closely, then fixing problems by asking for changes.
Voice agent
A voice agent is an AI system you talk to out loud that listens, decides what to do and answers in speech, often handling tasks like phone calls, bookings or support.
Voice cloning
Voice cloning is the use of AI to learn the sound of a specific person's voice from recordings and then generate new speech in that voice from any text.
VRAM
VRAM is the dedicated memory on a graphics card, and for AI it sets the limit on how large a model you can run locally, since the model's weights and working data must fit in it.

W

Web crawler
A web crawler is an automated program that visits web pages and follows links to read and collect their content, used by search engines and by AI companies gathering data to train or ground models.
Web scraping
Web scraping is using software to automatically copy content from websites at scale, and AI companies use it to gather text and data for training models and for answering questions.
Weight initialization
Weight initialization is how a neural network's parameters are set to starting values, usually small random numbers, before training begins, and it affects whether training works at all.
Word error rate
Word error rate (WER) is the standard score for speech-to-text accuracy: the share of words a system gets wrong compared with a correct transcript, counting substitutions, deletions and insertions, so lower is better.
World model
A world model is an AI system that learns how an environment behaves so it can predict what happens next and simulate it, such as generating interactive video, game scenes or robot training scenarios.

Z

Zero data retention
Zero data retention (ZDR) is an arrangement with an AI provider where your prompts and the model's outputs are not stored after the request is handled, usually approved per customer and limited to certain features.
Zero-day vulnerability
A zero-day vulnerability is a software flaw unknown to its maker, or known but not yet fixed, so defenders have had zero days to prepare a patch while attackers may already be using it.
Zero-knowledge proof
A zero-knowledge proof is a cryptographic method that lets one party prove a statement is true without revealing the secret behind it, such as proving a computation was done correctly.
Zero-shot learning
Zero-shot learning is when an AI model does a task it was never specifically trained or shown examples for, using only an instruction and what it already knows.