# AI glossary: plain-English definitions

> This glossary defines the key terms for using AI safely: AI hallucination, cognitive surrender, sycophancy, prompt, custom instructions, training on your conversations, placeholder, source click, pre-flight check, fresh chat, prompt library and spaced repetition.

Source: https://shaaz.online/guides/ai-glossary.html · Published 2026-10-04 · Updated 2026-10-05

The terms people meet when they start using ChatGPT, Claude and Gemini seriously, explained without jargon.

## AI hallucination

An answer from an AI that sounds confident but is wrong, invented or unsupported, such as a made-up fact, number, quote or source. The AI does not know it is wrong. See [how to spot AI hallucinations](https://shaaz.online/guides/how-to-spot-ai-hallucinations.html).

## Cognitive surrender

The tendency to accept AI output with minimal scrutiny, overriding your own intuition and reasoning. The term comes from Wharton researchers Steven Shaw and Gideon Nave.

## Sycophancy (AI)

The tendency of a chatbot to agree with you and flatter your ideas instead of challenging them. The Devil's Advocate Rule counters it by asking the AI to argue against you. See [why ChatGPT agrees with everything you say](https://shaaz.online/guides/stop-chatgpt-agreeing-with-everything.html).

## Prompt

The instruction or question you give an AI. Better prompts say who you are, what you need and why.

## Custom instructions

A settings section (called Custom Instructions or Personalisation in many apps) where you tell the AI about yourself once so it shapes every later answer.

## Training on your conversations

Many AI apps can use your conversations to train future models unless you switch that setting off. It is rarely on the front page, so check each tool you use.

## Placeholder

A stand-in such as "Client A" or "[Project Name]" that replaces a sensitive detail before you paste text into an AI, so you still get help without sharing private information.

## Source click

The habit of opening every source an AI gives you and finding the exact line that supports the claim. AI can produce links that look right but do not say what it claimed, or do not exist.

## Pre-flight check

A short routine you run before an AI answer leaves your screen and goes to a client, a customer or your bank.

## Fresh chat

Starting a new conversation instead of continuing a very long one. Each message makes the AI re-read the whole chat, so after many messages old details can blur.

## Prompt library

Your own collection of prompts that work, saved in one place you can reach in seconds.

## Spaced repetition

Reviewing something at increasing intervals so it sticks. The Pop Quiz in The 30-Day AI Habit uses it: right answers move to a slower review box and wrong answers return sooner.

## Sources

- [Shaw, S. and Nave, G.: Thinking, Fast, Slow, and Artificial: How AI is Reshaping Human Reasoning and the Rise of Cognitive Surrender (SSRN)](https://papers.ssrn.com/abstract=6097646)

## Frequently asked questions

**What is an AI hallucination?**
A confident AI answer that is wrong or made up, such as an invented fact, number, quote or source.

**What is cognitive surrender?**
Accepting AI output with minimal scrutiny, overriding your own thinking. The term is from Wharton researchers Shaw and Nave.