Below are answers to the most frequently asked questions about generative artificial intelligence at HAMK.

1. What are the key things to consider when using AI?

• AI is a support tool: AI should be seen as a tool that supports human activities rather than replacing them. It is best utilized for routine tasks and as an aid in brainstorming. The final responsibility and decision-making always rest with humans, and AI-generated results should be critically evaluated.
• Guiding students: The appropriate use of AI supports students’ learning and the development of workplace skills. It is essential to ensure that AI tools are used purposefully and that students receive guidance on their ethical and responsible use.
• Data protection: A good rule of thumb is that you should not feed artificial intelligence any information that you would not be able to post publicly online if the input data is used to train a language model. Special care should be taken with personal data, business and research information, confidential data, and licensed materials.More information on data classification can be found in this guide.

2. How do I get started with AI and where can I gain knowledge?

• The best way to start is by experimenting and learning through use. Ask for tips and share ideas with colleagues.
• Next, complete the AI module and competency badge available in the Staff Digital Pack (Intra).
• We also recommend following discussions and tips on the AI channel in the Teaching Technologies Teams group (HAMK staff) and participating in HAMK’s AI Coffee Sessions and training sessions.
• EduHouse offers a free training program on using ChatGPT: “ChatGPT – Work with AI.”

3. What are common applications of generative AI?

• ChatGPT: Developed by OpenAI, ChatGPT is a well-known natural language processing model capable of generating human-like text for various purposes.
• Gemini: Developed by Google, Gemini (formerly known as Bard) is a generative AI focused on natural language understanding and generation.
• Copilot in Edge (Bing AI Chat): Developed by Microsoft, this AI integrates with Microsoft 365 services, assisting with content creation and information retrieval.

4. Can I use these applications at HAMK?

• ChatGPT: You can log in with your HAMK Google credentials. By enabling “temporary chat” mode, conversations will not be stored in history, used for training, or retained in memory. You can find this setting in the chat menu under the language model selection.
• Gemini: You can log in with your HAMK Google credentials, ensuring that data usage for model training is disabled by default.
• Copilot in Edge: You can log into Edge’s Copilot using your HAMK account (username@hamk.fi), which provides access to a more secure version of Copilot.

5. What AI application can I use for image creation?

• Adobe Firefly: Developed by Adobe, Firefly is designed for creative content generation. It is integrated into Adobe Creative Cloud tools such as Adobe Express and Photoshop. Firefly’s training material is ethically sourced, using licensed images and models, making it a safe choice. It does not train its model based on user prompts.
• DALL-E: Developed by OpenAI, DALL-E generates images and can be used through ChatGPT or Copilot in Edge. Note that the free versions have a limited number of daily generations.

6. I want to use Microsoft Copilot. How do I proceed?

A separate M365 Copilot license (approximately €27/month) is required for use. The steps for obtaining it are:

1. Complete the AI Fundamentals competency badge (instructions on Intra, link).
2. Fill out the request form in Helpdesk. The form is located under “Applications and Licenses” as “Microsoft 365 Copilot License.”

7. I would like to access to a paid AI application. What should I do?

Submit a ticket to Helpdesk to start the needs assessment and procurement process. To acquire a paid AI application, you must first complete the AI competency badge (instructions on Intra, link). 

8. How do I use AI while respecting copyright laws?

The fundamental principle is that you should not upload someone else’s material into an AI application for analysis. Always check the rights of the materials you use and obtain necessary permissions before sharing or using AI-generated content. Read a more detailed explanation about AI and copyright. 

9. How much energy and water does AI use?

The energy consumption and water consumption of a single text response from AI are very small: producing a text response takes on average only a fraction of a watt-hour and only a few drops of water, which is comparable to normal search engine use or a short video streaming of a few seconds.

Creating images and especially videos consumes significantly more energy than producing text. Generating a single image typically consumes around 1–10 Wh of electricity, i.e. many times more than a single text response, and multiple iterations or high resolution increase consumption rapidly. Video generation is even more demanding because it combines image generation, temporal continuity and often high resolution, making it one of the most energy-intensive uses of generative AI. Therefore, the decisive factor in terms of overall consumption is how much and for what purpose images and videos are produced, not so much the individual generation.

The biggest environmental impacts arise from scale, i.e. the growth in the total number of AI workloads and especially image and video generation, which consume significantly more energy than text production.

At the same time, the energy efficiency of data centers and equipment is improving rapidly, which curbs consumption per work performed, and net water consumption is mostly low, as cooling water is largely returned to the environment.