Xerte Community Documentation

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Xerte Page-Specific AI Assistant Reference ======

This page contains individual page-specific **AI Assistant** documentation pages.

Largely, each AI Assistant is different and contains different fields tailored to the page in question. For example, the AI Assistant for the Quiz page differs significantly from the AI Assistant for the Flash Cards page. This page contains a detailed description and tips for each unique AI Assistant.

Note that common settings shared across each AI Assistant (including options for modulating difficulty, tone and style, or uploading files) will be covered only for page-specific advice. For a general overview, please see the [[assistants:ai_assistant_landing_page|AI Assistants Common Properties]] page.

===== Note on accuracy =====

Regardless of page, generated content should be reviewed before publication, especially when accuracy and alignment with source material are important. Models are prone to hallucination and making mistakes, the risk of which may be more or less present depending on the subject matter and the type of page being generated. Double-checking is a must!


----
====== Quiz: AI Assistant ======

The **AI Assistant** for the **Quiz** interactivity supports rapid quiz creation by generating questions and answer options from a defined subject.

===== General =====

Use the AI Assistant to quickly create quiz content for a specific topic.

You define the subject, the number of questions, the number of answer options per question, and the number of correct answers. The AI Assistant then generates quiz questions based on those settings.

===== Interface =====

{{:playground:ai_quiz_interface.jpg?400}}

===== Video Demo =====

https://xot.xerte.org.uk/preview_650#page1section11

{{url>https://xot.xerte.org.uk/preview_650#page1section11}}

Or view at [[https://xot.xerte.org.uk/preview_650#page1section11|Xerte AI Demos > Interactivity: Quiz]].

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The topic on which the quiz will be based. | AI input | Can be a single word, phrase, or full sentence. |
| **Number of Questions** | The number of questions to generate. | Numeric | Determines the total number of quiz items created. |
| **Answers per Question** | The number of answer options for each question. | Numeric | Determines how many answer choices are generated for each question. |
| **Number of Correct Answers** | The number of generated answers that should be valid correct answers. | Numeric | Determines whether generated questions use **single-choice** or **multiple-choice** selection. A value of **1** produces single-choice questions with radio buttons. A value greater than **1** produces multiple-choice questions, allowing more than one answer to be selected. |


----
====== Basic Text Generation: AI Assistant ======

Several Xerte pages use the same basic text-generation AI Assistant pattern.

===== Pages covered =====

This AI Assistant pattern applies to the following pages or page parts:

^ Area ^ Pages / availability ^
| Text pages | Columns, Plain Text |
| Navigators | Multi-Navigator, Multiple Perspectives, Tabbed Navigator+ |
| Interactivities and media pages | Documentation, 360 Image, Advanced Image, Audio Slideshow, Flickr Slideshow, Graphics and Sound, Image Sequence, Image Viewer, Morph Images |

Depending on the page, the AI Assistant may be available at top level, on sub-pages, or on individual content items such as columns, tabs, slides, or images.

===== General =====

The Basic Text Generation AI Assistant creates text in a simple structure based on a subject and optional additional guidance.

These pages have two notable fields:

  * **Subject**
  * **Additional Instructions**

The **Subject** field is mandatory and determines the main subject of the generated text. It can contain a word, phrase, or sentence. Although it is possible to provide detailed instructions in the Subject field, it is usually better to keep the subject brief and place extra guidance in **Additional Instructions**.

The **Additional Instructions** field is optional. It can be used to refine the output, add constraints, refer to uploaded source material, or explain the intended format.

For example, after uploading a PDF as context, you could write the following in **Additional Instructions**:

<code>
Create a summary of the end results of the research paper.
</code>

===== Advanced settings =====

In the advanced settings pane, the standard fields for reading level, education level, tone, and style are available.

The **Example Content** field lets you provide an example of the structure, wording, or format you want the generated text to follow.

For example:

<code>
Summary of [insert title of research paper here]

Main points:

Conclusion:
</code>

This encourages the AI Assistant to include the title, main points, and conclusion as separate parts of the generated response.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The main topic for the generated text. | AI input | Required. Can be a word, phrase, or full sentence. |
| **Additional Instructions** | Extra guidance for the generated text. | AI input | Optional. Use this for detail, formatting guidance, source-specific instructions, or constraints. |
| **Example Content** | An example of the desired output format or style. | AI input | Available in advanced settings. Helps guide structure and formatting. |
| **Reading Level** | Sets the intended reading level of the generated text. | AI setting | Shared AI Assistant setting. |
| **Education Level** | Sets the intended education level of the generated text. | AI setting | Shared AI Assistant setting. |
| **Tone and Style** | Sets the intended tone and style of the generated text. | AI setting | Shared AI Assistant setting. |

===== Best practice =====

Keep the **Subject** concise and use **Additional Instructions** for detail.

For larger or more specific tasks, provide source material where possible and tell the AI Assistant exactly what to focus on.

----
====== Orientation: AI Assistant ======

The **AI Assistant** for **Orientation** uses the existing learning object as context and attempts to generate an orientation page based on it.

===== General =====

The AI Assistant can create an orientation page that introduces the learner to the content of the learning object.

The generated orientation can include aims, objectives, a summary of the learning object, or guidance about how the learner should approach the material.

Please note that this is done based on the latest saved preview. To ensure the model is up to date when generating the orientation page, make sure to preview your learning object first!

===== Instructions field =====

The **Instructions** field lets you define the scope or focus of the generated orientation page. 

For example:

<code>
Focus only on the Flash Cards and Interactive Video sections.
</code>

This tells the AI Assistant to focus the orientation page on those parts of the learning object.

===== Best practice =====

When using the Instructions field, refer to specific page names, section names, topics, or terms from the learning object. This helps the AI Assistant identify the most relevant material.

If no instructions are provided, the AI Assistant attempts to create a general orientation page from the learning object. In larger learning objects, the result may focus on some sections and omit others because of size constraints. For larger projects, it is better to identify the pages, sections, or topics that should be prioritised.

The Instructions field can also accept general writing guidance, such as:

<code>
Keep the goals section short.
</code>

or:

<code>
Make the aims and objectives more detailed and focused on the cellular biology aspects.
</code>

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Instructions** | Guidance about what the generated orientation page should focus on. | AI input | Optional. Can refer to specific pages, sections, topics, or desired writing style. |

===== Note on accuracy =====

Review the generated orientation page to make sure it accurately reflects the learning object and does not omit essential material.

----

====== Button Sequence: AI Assistant ======

The **AI Assistant** for **Button Sequence** helps create sequence content by generating text on a subject and splitting it into several logical parts.

===== General =====

Use this AI Assistant to generate a block of text and divide it into sequence parts. A button sequence is, essentially, a sentence or piece of text underneath which a button can be found, and each button reveals the next segment of the content.

Normally, you'd have to split the text manually, add nodes, and then paste the parts in the desired order. With the assistant, this is done automatically.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The subject for the generated sequence text. | AI input | Required. Can be a word, phrase, or sentence. |
| **Sequence Parts** | The number of segments into which the generated text should be split. | Numeric | Each segment should have a logical stopping point. |
| **Sequence Length** | The approximate length of the generated text. | AI input / length setting | Can be expressed in words, such as **50 words**. |

===== Notes =====

The AI Assistant attempts to follow the requested sequence length, but may generate slightly more or less text where this improves clarity or keeps the sequence parts coherent.

----
====== Categories: AI Assistant ======

The **AI Assistant** for **Categories** can populate user-defined categories with suggested words or items.

===== General =====

Categories activities can be time-consuming to build manually. The AI Assistant can help by filling each category with suitable words or items, which can then be reviewed and edited.

You must define the categories yourself. The AI Assistant does not create the category names from a single subject. Instead, every category is treated as a separate 'subject' for which words are generated.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Categories** | The categories to populate. | AI input | Enter categories separated by commas. |
| **Words per Category** | The number of words or items to generate for each category. | Numeric | Determines how many items are added to each category. |

===== Using the Categories field =====

List the categories separated by commas.

For example:

<code>
Famous cities in France, Birds native to Europe, Musical instruments
</code>

The AI Assistant works best with clear, simple categories. A category can be a single word, a phrase, or a longer description, as long as commas are used only to separate categories.

===== Important note =====

Because commas are used as separators, avoid using commas inside a single category name. For example, //'Birds, Bees and Other Flying Critters'// would result in three categories: Birds - Bees - and Other Flying Critters.

----
====== Crossword Puzzle, Hangman and Wordsearch: AI Assistant ======

The **AI Assistant** for **Crossword Puzzle**, **Hangman**, and **Wordsearch** can quickly generate words based on a defined subject.

===== General =====

Use this AI Assistant to create a list of subject-specific words for games and word-based activities.

After defining the subject, specify the total number of words to generate.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The subject for the generated words. | AI input | Required. Should be specific enough to produce relevant words. |
| **Number of Words** | The number of words to generate. | Numeric | Determines the total number of generated words. |

===== Complexity and level =====

The standard AI Assistant options for reading level, education level, tone, and style are also be available.

Because these activities consist mainly of words, the subject and level settings should be chosen carefully.

For example, if the subject is:

<code>
Computational Neural Networks
</code>

but the reading and education levels are set to:
<code>
Beginner English
</code>
and
<code>
First Grade
</code>

the generated words may still be too complex because the subject itself does not lend itself well to simple explanations.

===== Best practice =====

Be specific about both the subject and the intended learner level. Review the generated words before publishing, as it's not uncommon for one or two to not fit the theme and/or be entirely hallucinated, depending on the model used.

----
====== Flash Cards: AI Assistant ======

The **AI Assistant** for **Flash Cards** can automatically generate flash cards from a subject.

===== General =====

Flash cards can be used for knowledge recall, vocabulary learning, language learning, revision, and concept checking.

The AI Assistant can generate cards in two main formats:

  * **Knowledge Recall** - the word appears first.
  * **Guess the Word** - the definition appears first.

===== Content order =====

The card format is controlled by the content order.

==== Knowledge Recall: word first ====

In this format:

  * Side 1 contains the word.
  * Side 1 may also include a hint.
  * Side 2 contains the definition or description.

==== Guess the Word: definition first ====

In this format:

  * Side 1 contains the definition.
  * Side 2 contains the word.
  * Hints are automatically disabled because the definition is already shown on the front side.

===== Language learning mode =====

By default, both sides of the flash card use the same language.

If **Language Learning Mode** is enabled:

  * the definition is generated in English, or in the language set at learning object level;
  * the word itself is generated in the language selected in the **Language Choice** field.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The subject for the flash cards. | AI input | Required. Can be a word, phrase, or sentence. |
| **Amount of Flashcards** | The number of flash cards to generate. | Numeric | Determines the total number of generated cards. |
| **Content Order** | Determines whether the word or definition appears first. | Selection | Use definition first for Guess the Word activities. |
| **Include Hints** | Adds hints to the front side of cards where applicable. | Yes/No | Available for Knowledge Recall cards. Disabled for Guess the Word cards. |
| **Language Learning Mode** | Generates cards for language learning. | Yes/No | Changes the word language according to the Language Choice field. |
| **Language Choice** | The target language for the word side of the flash card. | Selection / AI setting | Used when Language Learning Mode is enabled. |

===== Images in flash cards =====

Images are often paired with flash cards.

To use the Image Search and Help Tool with flash cards:

  - Open an individual card.
  - Add either the **Slide 1 Media** or **Slide 2 Media** optional property from the right sidebar.
  - Click the magic wand icon.
  - Enter a query.
  - Search one of the available image services or generate an image
  - Choose an image and select **Keep**.

The selected image link is automatically added to the media field of that particular flashcard.

----
====== Opinion: AI Assistant ======

The **AI Assistant** for **Opinion** can automatically generate opinion questions.

===== General =====

Use this AI Assistant to generate opinion questions based on a subject, categories, demographic, and scale.

The subject defines the overall topic. The categories define the areas within that topic that the questions should focus on.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The overall subject for the opinion questions. | AI input | Required. |
| **Categories** | The topic areas within the subject. | AI input | Required. Enter at least two categories separated by commas. |
| **Number of Questions** | The number of questions to generate for each category. | Numeric | Determines how many questions are created per category. |
| **Demographic** | The target audience for the survey or opinion activity. | AI input | Helps tailor the wording, relevance, and framing of questions. |
| **Scale** | The opinion scale used for the questions. | Selection / AI input | The default is the Likert scale. |

===== Using the Categories field =====

Categories can be words or short sentences. Each category must be separated by a comma.

Do not include commas inside a single category, as that would create two categories.

You must add at least two categories.

===== Best practice =====

Use the **Demographic** field to explain who the survey is for. For example:

<code>
Office workers with higher education between the ages of 35 and 60 working in <Insert City Name>
</code>

This helps the AI Assistant produce questions that are more relevant to the intended audience. Do note that even with detailed demographic descriptions the questions might turn out less relevant, or otherwise not applicable.

----
====== Answer X of Y: AI Assistant ======

The **AI Assistant** for **Answer X of Y** can generate the question and possible answers for X of Y style questions.

===== General =====

Use this AI Assistant when the correct answers come from a finite, countable set.

The AI Assistant can also generate common abbreviations or alternative spellings where appropriate.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The finite and countable subject for the question. | AI input | Required. Can be a word, phrase, or sentence, but should refer to a countable set. |
| **Number of Options** | The number of possible answers to generate. | Numeric | Determines how many answer options are created. |

===== Best practice =====

Choose a subject with a clear set of possible answers.

For example, a subject such as:

<code>
Countries in Scandinavia
</code>

is more suitable than a broad or open-ended subject such as:

<code>
Important countries
</code>

----
====== Interactive Video: AI Assistant ======

The **AI Assistant** for **Interactive Video** can generate interactive overlay content for a video.

===== General =====

The Interactive Video page always consists of two sub-pages: the media panel, which hosts the video, and the overlay panel, which handles the content over the video (text and questions).

The AI Assistant is available on the **overlay panel** sub-page.

The overlay panel includes a video input field, available in the advanced settings, where you can upload a video or enter a PeerTube, YouTube, or Vimeo link, as well as options to choose to generate a multiple choice question or text block, synced to the respective section of the video.

The video in the overlay panel should match the video used in the media panel. By default, the overlay panel inherits the link used in the media panel. If you change it, make sure that change is reflected in both fields!



===== Object types =====

The AI Assistant can generate two main types of content:

  * **Text objects**
  * **Multiple choice questions**

Use the **Object Type** drop-down menu to select the type of object to generate.

===== Generate text object =====

Text objects are useful for start cards, end cards, fun facts, reminders, summaries, or other video overlay text.

The **Text Object Details** field lets you describe the type of text objects to generate.

Separate each requested text object with a comma.

For example:

<code>
Start Card, Fun Fact, End Card
</code>

This should generate an introductory card, a fun fact box based on the video, and an ending card.

The wording does not need to be exact. You can describe the requested text objects in different ways.

For example:

<code>
3 text objects, each describing the philosophy, the research results and the societal impact respectively
</code>

This should work where the video transcript contains corresponding content.

===== Generate multiple choice questions =====

Multiple choice questions can be used to refresh knowledge, test understanding, or add interactive checkpoints to the video.

Set the number of questions and the number of possible answers.

The AI Assistant generates questions based on the subject provided and relevant content found in the video.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Video Input** | The video used as context for generation. | Media input / URL | Can be an uploaded video or a PeerTube, YouTube, or Vimeo link. Should match the media panel video. |
| **Object Type** | Selects what type of overlay content to generate. | Selection | Choose text objects or multiple choice questions. |
| **Text Object Details** | Describes the text objects to generate. | AI input | Separate multiple requested text objects with commas. |
| **Number of Multiple Choice Questions** | The number of multiple choice questions to generate. | Numeric | Used when Object Type is set to multiple choice questions. |
| **Number of Answers** | The number of answer options for each multiple choice question. | Numeric | Used when generating multiple choice questions. |
| **Subject** | The topic the generated objects or questions should focus on. | AI input | Used with video content to guide generation. |

----
====== Text Correction: AI Assistant ======

The **AI Assistant** for **Text Correction** helps generate two texts on a given subject at the same time.

===== General =====

This AI Assistant is designed for activities where learners compare or correct texts.

It can generate a correct text and a second text containing deliberate mistakes.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The subject for the generated texts. | AI input | Required. |
| **Mistake Type(s)** | The types of mistakes to include in the second text. | AI input | Can include grammar, spelling, prepositions, word order, or other language-related mistake types. |
| **Text Length** | The approximate desired length of the generated texts. | AI input / length setting | Best expressed as an approximate number of words. |

===== Mistake Type(s) field =====

The **Mistake Type(s)** field is the most important field for controlling the second text.

It can include one type of mistake, several types, or guidance about frequency.

Examples:

<code>
Prepositions
</code>

<code>
Grammar, spelling, and word order
</code>

<code>
Grammar, very frequent mistakes
</code>



----
====== Interactive Text: AI Assistant ======

The **AI Assistant** for **Interactive Text** supports two tasks:

  * generating a sentence;
  * marking words by groups.

The task is selected from the **Task Type** drop-down menu.

===== Task type: Generate a Sentence =====

When the task is set to **Generate a Sentence**, the AI Assistant creates text on the given subject.

The most important field here is **Sentence Length**, which controls the approximate number of words in the generated sentence.

===== Task type: Mark Words by Groups =====

When the task is set to **Mark Words by Groups**, the AI Assistant uses the **Word Groups** field to identify the types of words that should be marked.

Examples of word groups include:

  * verbs
  * nouns
  * prepositions
  * subjects
  * colours
  * animals

Multiple word groups can be entered, but avoid overlapping categories where possible.

For example, marking both **colours** and **adjectives** may cause ambiguity because many colour words can also function as adjectives.

===== Passage field =====

The **Passage** field automatically copies text from the **Interactive Text** field.

This is the text that the AI Assistant will make interactive. The passage can be edited in the AI Assistant window before generation.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Task Type** | Selects the task the AI Assistant should perform. | Selection | Options include Generate a Sentence and Mark Words by Groups. |
| **Subject** | The subject for generated text. | AI input | Used when generating a sentence. |
| **Sentence Length** | Approximate length of the generated sentence. | Numeric / length setting | Usually expressed as a number of words. |
| **Word Groups** | The types of words to mark. | AI input | Used when marking words by groups. Avoid overlapping groups where possible. |
| **Passage** | The text to make interactive. | AI input / copied text | Automatically copies text from the Interactive Text field. Can be edited before generation. |

----
====== Simulating Question: AI Assistant ======

The **AI Assistant** for **Simulating Question** generates a thought-provoking question and supporting feedback.

===== General =====

Use this AI Assistant to create questions that encourage further thinking about a subject.

The generated feedback can include information, prompts, and suggestions about how the question can be used for deeper reflection or discussion.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The subject for the simulated question. | AI input | Required. |
| **Reading Level** | Sets the intended reading level of the generated question and feedback. | AI setting | Shared AI Assistant setting. |
| **Education Level** | Sets the intended education level of the generated question and feedback. | AI setting | Shared AI Assistant setting. |
| **Tone and Style** | Sets the tone and style of the generated question and feedback. | AI setting | Shared AI Assistant setting. |

===== Best practice =====

Use the advanced AI Assistant settings for language proficiency, reading level, and education level. These settings help control the complexity and suitability of the generated question which is especially important for this page.
----
====== Sortable Grid: AI Assistant ======

The **AI Assistant** for **Sortable Grid** can create grids based on a given subject.

===== General =====

Use this AI Assistant to generate a table-like grid of observations or entries.

The subject defines what the rows are about, while the columns define the characteristics that should be included.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The subject for the grid. | AI input | Required. |
| **Columns** | The desired columns or characteristics of the subject. | AI input | Enter the columns that should appear in the grid. |
| **Number of Rows** | The number of observations or entries to generate. | Numeric | Determines how many rows are added to the grid. |

===== Example =====

If the subject is:

<code>
Birds
</code>

the Columns field could include:

<code>
wing span, colour, species
</code>

The AI Assistant will then generate rows using those characteristics.

----
====== Dialogue: AI Assistant ======

The **AI Assistant** for **Dialogue** helps create pairs of questions and answers.

===== General =====

The AI Assistant generates:

  * questions;
  * answers;
  * possible answer variations;
  * hints.

This is useful when several different answers may be valid.

For example, when asking learners to explain an abbreviation, the AI Assistant can generate possible alternative phrasings that should still be accepted as correct.

===== Delimiter note =====

When unsure, set the delimiter to a comma.

If a different delimiter is needed, add the delimiter optional property and define it there. The AI Assistant settings window will inherit the same value.

This is important because the AI Assistant needs to know how to separate provided terms and generated answer variations.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The subject for the dialogue items. | AI input | Required. |
| **Delimiter** | The character used to separate terms or answer variations. | Text / setting | Use a comma unless another delimiter is required. |
| **Questions and Answers** | Generated dialogue pairs. | Generated content | The AI Assistant creates question and answer pairs. |
| **Answer Variations** | Alternative valid responses. | Generated content | Useful where more than one phrasing may be correct. |
| **Hints** | Hints for each generated question. | Generated content | Helps learners answer the question. |

----
====== Gap Fill: AI Assistant ======

The **AI Assistant** for **Gap Fill** supports two tasks:

  * generating a sentence;
  * marking words as blanks.

The task is selected from the **Task Type** field.

===== Task type: Generate a Sentence =====

When the task is set to **Generate a Sentence**, the AI Assistant generates text that is placed in the **Gap Fill Text** field.

===== Task type: Mark Words as Blanks =====

After generating text, or after writing your own text, switch the task type to **Mark Words as Blanks**.

The **Word Groups** field lets you specify the types of words that should become blanks.

Examples include:

  * nouns
  * adjectives
  * verbs
  * colours
  * animals
  * objects

===== Passage field =====

The **Passage** field inherits text from the **Gap Fill Text** field, including the last generated sentence.

The passage can be edited in the AI Assistant window before generation or left unchanged.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Task Type** | Selects the task the AI Assistant should perform. | Selection | Options include Generate a Sentence and Mark Words as Blanks. |
| **Subject** | The subject for generated text. | AI input | Used when generating a sentence. |
| **Word Groups** | The types of words to convert into blanks. | AI input | Used when marking words as blanks. |
| **Gap Fill Text** | The generated or manually written text for the activity. | Page content field | Can be populated by the Generate a Sentence task. |
| **Passage** | The text to process when marking blanks. | AI input / copied text | Inherits content from the Gap Fill Text field. Can be edited before generation. |

----
====== Multiple Choice Question: AI Assistant ======

The **AI Assistant** for **Multiple Choice Question** generates answer options for an existing question.

===== General =====

The main **Question** field must already contain the question.

The AI Assistant then generates possible answers for that question.

===== Important limitation =====

The AI Assistant currently generates answers for the question provided in the main **Question** field.

It does not currently generate both the question and the answers from a subject in one 'go'.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Question** | The question for which answers should be generated. | Page content field | Must be provided before using the AI Assistant. |
| **Number of Answers** | The number of generated answers. | Numeric | Determines how many answer options are created. The minimum is 1. |
| **Number of Correct Answers** | The number of correct answers. | Numeric | Determines whether generated questions use single-choice or multiple-choice selection. A value of 1 produces single-choice questions with radio buttons. A value greater than 1 produces multiple-choice questions, allowing more than one answer to be selected |

===== Number of Answers & Correct Answers =====

The **Number of Answers** field determines how many answer options are generated.

If set to **1**, the AI Assistant generates only what it identifies as the correct answer and distractors can be added manually.

Higher values generate additional answer options, including distractors.

The number of correct answers determines the interactivity, and this will be set automatically. You do not have to change the 'Question Type' property after generation.

----
====== Inventory: AI Assistant ======

The **AI Assistant** for **Inventory** creates questions and multiple possible answers that categorise learners or respondents into defined classes.

===== General =====

Use this AI Assistant to generate inventory-style surveys or quizzes.

The AI Assistant creates:

  * questions;
  * answer options;
  * result categories;
  * interpretation text for the end user.

===== Classes field =====

The **Classes** field is the most important field.

It should contain the groups, categories, or outcomes into which users will be placed after completing the inventory.

Classes should be separated by commas.

The classes can be any grouping that makes sense in the context of the subject.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The subject of the inventory. | AI input | Required. |
| **Classes** | The end groups or categories users can be placed into. | AI input | Separate classes with commas. |
| **Questions** | Generated inventory questions. | Generated content | Created by the AI Assistant. |
| **Answer Options** | Generated answer options for each question. | Generated content | Used to categorise respondents into classes. |
| **Result Interpretation** | Generated interpretation of the user’s result. | Generated content | Explains the outcome to the end user. |

===== Best practice =====

Define clear and distinct classes. Overlapping classes may make the generated questions and interpretations less reliable. In general, this page is complex and contains a lot of sub-nodes, as well as weights which are assigned to each answer of each question. It's therefore worth it to double check things in detail, especially if you mean to make use of the graphs at the end of the inventory.

----
====== Matching Texts: AI Assistant ======

The **AI Assistant** for **Matching Texts** creates sentence pairs that learners must match.

===== General =====

The AI Assistant generates statements split into two parts. The learner then connects the matching parts.

A hint is also generated for each sentence pair.

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Subject** | The subject for the generated matching texts. | AI input | Required. |
| **Sentence Length** | Approximate length of each sentence part. | Numeric / length setting | Usually expressed as a number of words. |
| **Number of Sentences** | The number of sentence pairs to generate. | Numeric | Determines how many matching pairs are created. |
| **Hints** | Hints for each generated sentence pair. | Generated content | Helps learners identify the correct match. |

===== Notes =====

The **Sentence Length** field approximately determines the length of each part of the sentence.

The **Number of Sentences** field determines how many sentence pairs are generated at one time.

----
====== Modify Exercise: AI Assistant ======

The **AI Assistant** for **Modify Exercise** can either generate new text or alter existing text.

===== General =====

The AI Assistant has two tasks, selectable from the **Task Type** drop-down menu:

  * **Generate Text**
  * **Alter Text**

===== Task type: Generate Text =====

When the task is set to **Generate Text**, the AI Assistant creates a paragraph according to your specifications.

The generated paragraph is placed in the **Initial Text** field.

===== Task type: Alter Text =====

When the task is set to **Alter Text**, the AI Assistant takes existing text from the **Initial Text** field and places it into the **Text to Modify** field in the AI Assistant settings.

The AI Assistant then alters the text according to your specifications and places the result in the **Text** field in the model answer group.

The source text can be text generated by the AI Assistant or text entered manually.

===== Text Specifications field =====

The **Text Specifications** field controls how the text should be created or altered.

Examples:

<code>
Switch to past tense.
</code>

<code>
Replace nouns with diminutives.
</code>

<code>
Make the paragraph more formal.
</code>

===== AI Assistant properties =====

^ Property ^ Description ^ Type ^ Additional information ^
| **Task Type** | Selects whether to generate text or alter existing text. | Selection | Options include Generate Text and Alter Text. |
| **Subject** | The subject for generated text. | AI input | Used when generating text. |
| **Initial Text** | The starting text for the exercise. | Page content field | Can be generated by the AI Assistant or entered manually. |
| **Text to Modify** | The text that will be altered. | AI input / copied text | Inherits text from Initial Text when Alter Text is selected. |
| **Text Specifications** | Instructions for how the text should be generated or altered. | AI input | Use this field for precise transformation instructions. |
| **Text** | The altered model answer text. | Generated content / page field | The altered result is placed in the model answer group. |

===== Best practice =====

Use clear, concrete instructions in **Text Specifications**. The more precise the transformation, the more predictable the generated result.