| AI chatbot |
An AI chatbot is a tool designed to simulate human-like conversations. You can think of it as a virtual assistant that you interact with. You can ask it questions or give it commands, and it responds conversationally. At the heart of an AI chatbot is a large language model, like GPT-4o, Llama, Claude, or others. |
| Bias |
AI chatbots are powered by large language models that may reflect biases present in their training data. As a result, the outputs produced may be biased.
To get more diverse responses from the AI chatbot, encourage the chatbot to provide information from different perspectives and cultures.
|
| Generative AI |
Generative AI refers to a type of application, such as ChatGPT, designed to use a variety of machine learning algorithms to create new content such as text, music, images that mimics human creation. |
| Hallucinations |
Hallucinations refer to where large language models produces content that is entirely fabricated.
Please verify the accuracy of the answers generated and remember that you are responsible for any content you create or share, including AI-generated content.
|
| Human-in-the-loop |
Human-in-the-loop is an approach to AI that integrates human input into the development of AI systems which may involve providing ethical oversight, correcting errors, and helping in training the AI. |
| Jailbreaking |
Jailbreaking, in the context of an AI tool such as an AI chatbot, refers to a technique used to modify or manipulate the chatbot to bypass built-in guardrails, controls, or functionalities. This may involve exploiting vulnerabilities to allow the AI chatbot to perform actions it was originally restricted from, such as accessing unauthorized data or operating outside of its designed ethical guidelines and constraints. |
| Large Language Model (LLM) |
An LLM is a type of AI system designed to understand and generate human language based on the text input it receives. These models are “large” because they are trained on vast amounts of data, allowing them to learn complex patterns of language usage, grammar, and information across diverse topics.
You will often interact with an LLM through an AI chatbot.
|
| Multimodal AI |
Multimodal AI, or Multimodal LLM, refers to advanced AI models that are designed to process, interpret and interact with a variety of input formats such as text, images, audio and video. |
| Prompt |
A prompt is how you ask an AI tool, for example an AI chatbot, to perform a task for you like creating, summarizing, or researching a topic. It is the message you type or speak into an AI chatbot. The chatbot then processes this prompt and generates a response based on the instructions you’ve provided.
|
| Prompt-a-thon |
A prompt-a-thon is a collaborative, hands-on session where participants will explore generative AI, learn how to interact with generative AI, and discover how generative AI could play a part in their personal or professional lives. |
| Prompt injection |
Prompt injection is a technique where a user intentionally inputs crafted prompts to manipulate an AI tool, such as an AI chatbot, often to make it behave in unintended ways by exploiting its processing logic. This can lead to the AI bypassing restrictions, accessing unauthorized functionalities, or producing inappropriate responses. Unlike jailbreaking, which involves modifying an AI chatbot to circumvent built-in safeguards, prompt injection disguises harmful instructions as harmless inputs, subtly manipulating the chatbot’s behavior. |
| Retrieval-Augmented Generation (RAG) |
Retrieval-Augmented Generation is a method designed to enhance the content generated by a Large Language Model. It works by combining information retrieval with content generation, allowing the tool to access up-to-date information or specialized datasets. This ensures the generated content is more relevant and accurate. |
| System prompt |
A system prompt refers to the predefined text or message that defines the desired personality of an agent. This includes any rules or guidelines which it follows when generating a response. It is not all AI chatbots that allow you to set a system prompt as it is often defined in the background and it is not something you will see when interacting with an AI chatbot. |
| Token |
Tokens are single units that serve as building blocks for a sentence or document (such as a words or characters). They are a measurement tool used to assess the length of chat inputs and outputs, defining the amount of work a LLM must perform.
The process of breaking text into tokens is known as tokenization.
|