Artificial intelligence supports you in developing digital learning content in Knowledgeworker. Let it generate appropriate structures, text sections, or images to create e-learning content faster than ever before. Afterwards, you can translate the courses into the languages required by your company with just a few clicks. Below, we will address frequently asked questions about AI.
What AI-based assistant functions does Knowledgeworker offer?
Knowledgeworker enables the following with the help of artificial intelligence:
- Creation of structures for e-learning courses
- Generation of texts
- Generation of relevant images
- Translation of learning content
The decision of whether to use all or only selected assistant functions is up to you.
Are my data safe when using the AI assistants?
Knowledgeworker utilizes services from third-party providers ("AI services") to provide AI assistant functions. Certain information needs to be transmitted to these AI services for them to perform their functions. For example, the text in the source language is transmitted to an AI translation service to generate a text in the target language. In this section, you will learn about the information that is transmitted to the AI services.
Text and image generation
To generate a course structure in Knowledgeworker, the title and description of the course are transmitted.
For generating appropriate text snippets in the Knowledgeworker Create editor, a chat view is available. All messages in the chat are transmitted to the AI service. The responses are stored exclusively in Knowledgeworker Create and only if they have been actively incorporated into the content.
For generating an image, only the author's description for the intended image is transmitted. The image is stored exclusively in Knowledgeworker, not in the AI service.
This means that you always have the independent decision-making power regarding which data you transmit and which you do not. The integrated AI services in Knowledgeworker ensure that the transmitted data is not used for training the respective AI. However, you should avoid transmitting confidential information to the AI services.
When using the integrated AI translation, the content to be translated is fully transmitted to the translation service in the source language. We have a contractual agreement with the AI service that the transmitted data will not be stored or used for training the AI. Additionally, we have an order processing agreement with the AI service in case the transmitted content contains personal data.
There is no transmission of personal data to the AI services since the requests to the AI services are sent exclusively by Knowledgeworker, not directly or on behalf of the user. This does not apply if the author directly includes personal data in the request (prompt).
In general, we recommend enabling authors to use the integrated AI-based assistant functions in Knowledgeworker, as this ensures better protection of both personal data and transmitted information compared to independent use of various tools by your authors. Additionally, the workflow is significantly more efficient since content does not need to be copied between different external tools.
Security measures of the currently used services
- https://openai.com/security (text and image generation)
- https://www.deepl.com/en/pro-data-security/ (translation)
Do I need to train my authors to use the AI assistants?
The AI functionalities are generally self-explanatory, so no training is required for their application.
However, we recommend sensitizing your authors to the responsible use of the AI assistants before granting access to them.
- Confidential data such as product descriptions, customer data, or any other company internal information should not be entered into the chat fields for text generation (prompts).
- AI-generated content is based on statistical probabilities. Although they have now reached a relatively good quality, one should never blindly rely on their accuracy.
- AI services can also generate discriminatory, defamatory, or misleading content as the algorithms incorporate societal biases (cultural, racial, religious, ideological biases) into their models.
- The issue of copyright for AI-generated content is still under discussion. AI services generate texts and images based on the information they have been trained on. It is therefore possible that the AI reproduces protected symbols, existing text structures, slightly modified text excerpts, or even quotations. Authors should individually assess whether certain texts contain copyrighted parts.