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 AI Assistants Common Properties page.

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!

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

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.

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

Or view at Xerte AI Demos > Interactivity: Quiz.

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

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

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.

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:

Create a summary of the end results of the research paper.

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:

Summary of [insert title of research paper here]

Main points:

Conclusion:

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

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.

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

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!

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

For example:

Focus only on the Flash Cards and Interactive Video sections.

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

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:

Keep the goals section short.

or:

Make the aims and objectives more detailed and focused on the cellular biology aspects.

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

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

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.

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.

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

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.

List the categories separated by commas.

For example:

Famous cities in France, Birds native to Europe, Musical instruments

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.

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.

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

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.

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:

Computational Neural Networks

but the reading and education levels are set to:

Beginner English

and

First Grade

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

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.

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

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.

The card format is controlled by the content order.

In this format:

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

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.

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.

Images are often paired with flash cards.

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

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

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

The AI Assistant for Opinion can automatically generate opinion questions.

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.

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.

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

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

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.

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

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.

Choose a subject with a clear set of possible answers.

For example, a subject such as:

Countries in Scandinavia

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

Important countries

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

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!

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.

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:

Start Card, Fun Fact, End Card

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:

3 text objects, each describing the philosophy, the research results and the societal impact respectively

This should work where the video transcript contains corresponding content.

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.

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

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.

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:

Prepositions

Grammar, spelling, and word order

Grammar, very frequent mistakes

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.

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.

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.

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.

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

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.

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.

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

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.

If the subject is:

Birds

the Columns field could include:

wing span, colour, species

The AI Assistant will then generate rows using those characteristics.

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

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.

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.

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.

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

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

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.

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

The main Question field must already contain the question.

The AI Assistant then generates possible answers for that question.

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'.

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.

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

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.

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.

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.

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

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.

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.

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

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

• Generate Text
• Alter 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.

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.

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

Examples:

Switch to past tense.

Replace nouns with diminutives.

Make the paragraph more formal.

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