Dr. Ayse Ozturk

A Brief Video Summary of the Article Below (by Notebook LM)

Let’s start with a video summary of the article below to show an example of AI-based transformations we can use for teaching purposes. The video below was created using NotebookLM and provides a brief summary of the main points in the article.

AI is an Enigma to Everyone

Generative AI is a black box, and no one, including the creators of AI, truly understands the full extent of what it can or can’t do. That’s why most of us are experimenting in real time. We are not and should not be supposed to use it everywhere, but rather carefully identify the situations where its benefits and relevance outweigh the risks and costs. Here are some references on the enigma of generative AI tools:

• “You don’t really know what (these AI models are) capable of until they’re deployed to a million people.  You can test ahead of time. You can have your researchers bash against them… But the hard truth is that there’s no way to be sure.” – Dario Amodei, 2025.
• “Because (LLMs) are so complex, even their creators don’t fully understand how they work.” –  Zeynep Tufekci, NYT, 2025.

It’s All About Experimenting with AI:

Since AI is such a big enigma, users from all aspects of professions, from creators to businesspeople, educators, entrepreneurs, office staff, and students, are scrambling to understand and evaluate its use cases. This is essentially a massive collaborative experiment in which we are all taking part. The creators introduce this new technology, along with its benefits and risks, for us to discover. It is important to approach and use these tools as consciously as possible. Since there are no established norms, we can only explore them as they are being launched. The best way to get started with AI is to open a tab on one of the well-known foundation models (i.e., a large language model such as ChatGPT, Gemini, or Claude, trained on extensive datasets to perform a wide range of tasks), and start chatting with it or prompting it.   

What Can Educators Do?

Here are some ideas on how to integrate AI into teaching to enhance student learning:

1. First Stages – Brainstorming

Best forms of integrating AI in classroom learning:

• Improve the course content: You can upload your existing course materials (not copyrighted ones, although possibly they are already trained on those!) into AI and ask it to enhance the content (e.g., with additional examples, current examples, other relevant concepts, visuals, etc.).
• Create more tailored and relevant examples for your topics
• Develop a variety of learning exercises (e.g., ask AI to create role-playing games, application of concepts on real-life situations, etc.)

2. Second – Hands-on Learning, Generating Creative Content, Operational Efficiency

• Ask AI to generate multiple modes of learning
• AI can basically turn anything into anything else. Upload texts to create infographics, videos, storybooks (e.g., Gemini storybook), web interfaces, etc., to use in class.
• Example on Gemini:
  • Gemini: Interactive quiz, infographic, etc.
  • Create Gemini Gems (this is similar to Custom GPTs, Claude Artifacts, Boodlebox custom bots)
  • Connect to Apps —> Google Docs create a new slide. Open a PDF and chat inside.
• Other tools: E.g., create presentations, using Gamma, GenSpark.ai, Skywork.ai, etc.
• Boodlebox: Good for collaborative exercises. Students can chat with each other as well as with the chatbot within a given task.
• NotebookLM: It can help students create their own practice questions, quizzes, flashcards, study guides, lecture podcasts, and even video lectures.

3. Later on – Advanced Simulations-Learning Games and Trying Out Other Tools

• Develop simulations and games tailored for the topic, such as decision making, role playing, visualization, escape room, family feud type of games. See examples on this page.
• Develop actual business tools: E.g., CRM Portal on DeepAgent: https://crmportal.abacusai.app/dashboard
• Comet browser: Create a mini case on the fly from an article. See the right side of the image below. Comet browser can create the case on the page with one prompt. Copy-paste to a file and share with students to discuss.

• Turn news articles into many forms of learning tools:
  • Example: I read this article on WSJ and turned it into assignments:
  • Assignment 1: https://drayseozturk.org/ai-team-of-rivals-approach/
  • Assignment 2: https://marketingteamofrivals.abacusai.app/dashboard
• Other creative uses:
  • Creating interactive flashcards using Kuse.ai
  • Writing an article (SEOwriting)
  • Writing a textbook (GenSpark)

Tips for Efficiency with AI Tools:

• For more efficiency, you can chat with AI during this brainstorming phase using the voice mode available in many frontier chatbots. It helps save time since you can chat during your daily walk, commute, while doing the dishes, getting ready for school, etc. The chatbot transcribes these oral communications, so you can also refer back to and copy the dictated transcript if needed. You can further clear it, summarize/synthesize the key points, and follow up in a written format later.
• Another efficiency tip is to turn your files into audio. NotebookLM can generate podcasts from uploaded files. But to listen on the fly, there are tools like Natural Readers or Microsoft Edge browser’s Read Aloud or Google Gemini’s audio converter on Google Docs. You can upload your document or paper and hit the play button, for example, during your commute to work. There are also paid apps like Speechify that turn text into speech.  
• You don’t need to use so many AI tools out there, but even if you use one, knowing how to use it well matters. For example, using the “Projects” function in ChatGPT helps increase efficiency. You can easily find your ongoing chats there and do not need to provide the same detailed prompts to the bot each time, since it already has the prompt from earlier. You can simply continue the same chat whenever you return to where you left off.

Other ways to help students

• Give them information about the free AI tools available for students. See the “Free Offers Worth Considering” section on this page for AI for Educators.
• Provide them with helpful updates, such as Gemini (free for students), which can generate professional headshots from a random selfie.
• Make students aware of AI’s limitations (hallucinations and biases, etc.). Remind them to check for accuracy (especially if accuracy is critical for the task). Although not foolproof, these tools have “fact-checking” features: E.g., GenSpark, Gemini.
• Try different approaches for the exercise: Students can start with an AI-written draft and improve it with their own inputs, or they can start with their own drafts and then get feedback from AI to finalize them.
• For easiest AI access: Use Comet (free version available) or Google AI Mode

Limitations and Challenges

Privacy and security are a big challenge. AI companies are known for not being transparent in their product development and data gathering processes. The best way to address this is to use a limited number of free AI tools that do not require a sign-in. Or you can ask students to create a backup email account only to be used for AI tools. Alternatively, if your institution offers a secure AI tool that maintains data confidentiality, you can encourage its use.

Relatedly, it is important to remind students not to upload any sensitive information to these AI tools. The only way to employ this approach is limited to personal ethics. At least warning students about which files would be fine to upload can help.

Another challenge is that some students do not want to take part in these AI exercises. You can expect some resistance against AI. This is especially evident when the AI tool requires a sign-in or if the free AI tool requests credit card information to be charged after the free trial ends. These instances lead the student to be anxious about participating in the AI-led exercise. AI is not like social media that immediately appeals to youth in certain aspects. The instructor then needs to create backup options for the students who are not willing to participate in the AI work. However, a new challenge arises then because of the lack of or imbalance of standards for AI expectations.

One last, and perhaps the most critical challenge, is the current path of AI progress toward increasing connectivity with other apps and all parts of the digital world. Many foundation models, such as ChatGPT, Claude, and Gemini, have the capability to connect to our apps on our devices. The AI agents can send emails, edit our files, sign up, and make changes within our online accounts with just a prompt. The new technical terms that we keep hearing more and more, such as API (Application Programming Interface) and MCP (Model Context Protocol), are simply doors for software programs to communicate and connect with each other. As AI becomes more complex and intrusive in every aspect of our digital lives, it becomes increasingly challenging to teach students these complex aspects and expect them to connect AI to all their files and the digital world. Nothing is likely to stay independent from AI anymore. Some students will rightfully not want to participate in the increasing integration needs of AI tools. AI will still make mistakes and can ruin important digital work, but more importantly, how far of an AI integration we can reasonably expect from students, and still maintain the same standard for everyone, will likely remain a significant challenge.

For Further resources, check the page “AI for Educators.”