1615 week3 Week 3 – July 15 to 21: What AIED Tools

1615 week3

Week 3 – July 15 to 21: What AIED Tools are Out There? Generative AI

Question1(150 words)

Icebreaker: Explore ChatGPT

Watch this video that introduces ChatGPT (2mins):

https://youtu.be/EPO9zZgyh7o?si=qAOkiFc9Uy5O4OLl

Then also watch this: “ChatGPT: The Pros and Cons of AI in Education and Beyond” (8 mins):
https://www.youtube.com/watch?v=s7kvmcxWbm8 

Explore ChatGPT yourself:
https://chatgpt.com/ 
(The free version uses ChatGPT 3.5 offered by Open AI which does not require sign-up. Creating an Open AI account, though, will enable you to have free access to GPT-4o which is more powerful GPT version with voice messaging ability.)

If you’re unable to access ChatGPT due to the restrictions or don’t feel comfortable creating an account post the prompt you’d like to ask and we’d be happy to send you the results!

What is your general impression from trying out ChatGPT? Did anything surprise you? How did you find the experience?

Question2(150 words)

Large language models (LLMs) can provide both accurate and inaccurate information, depending on their training data and underlying architecture. This phenomenon, sometimes referred to as the “jagged frontier,” presents a challenge in distinguishing valid information from misinformation.

(1) Given that AI outputs are not inherently “hallucinations” but can still be unreliable, how can professionals across fields—such as educators, designers, and administrators—work to inform others about these limitations and help them navigate the landscape of AI-generated information? 

(2) How do we mitigate the errors of AI in an educational setting? There is no such thing as a foolproof method of verification, so how can we use this to our benefit?

(3) Aditionally, how do we think critically about what an AI system generates? 

Looking forward to your thoughts!

Give some your thoughts to this discussion(50 words)

I am just thinking back to how quickly the staff at our school became acquainted with various AI tools, but there was really no conversation about its limitations or any major ethical considerations at the time (besides perhaps the issue of plagiarism). I think with time, we’ve come to realize that the outputs of AI, especially of LLMs like ChatGPT can be unreliable very often. One way to keep professionals informed about its limitations may be to have workshops and seminars explaining how these GPTs work, but not just theoretically, rather it would be good for these training workshops to include practical examples as well.

As for mitigating errors, it is always good to cross-reference any material generated by AI with the past experiences of professionals. Especially in an educational context, most of the times AI is used to generate resources for teaching, lesson plans, or assessments; it is always good to check with other educators and get their feedback on what they think about it. In addition using your own past experience and granting it greater value and credibility in judging the output of AI is also important.

Question3(150 words)

Deep Fakes are super-realistic, AI-edited images or videos that can make people look like they’re doing or saying things they never did. For instance, in January, an AI-generated Deep Fake photo of Taylor Swift caused significant uproar and led to legal action. Imagine a fake video of your favorite celebrity doing something crazy going viral—sounds wild, right? Systems have put “guardrails in place to help prevent the spread of negative usage like this, but as mentioned in the video they are easily bypassed. To delve into the dangers of deepfakes use the following prompts to guide your discussion:

(1) Ethics Check: When do you think is it okay (or not okay) to use AI to create these kinds of images or videos?
(2) Safety First: What should schools, companies, and governments do to keep people safe from fake content?
(3) Getting Informed: How can we teach everyone about the risks and realities of Deep Fakes?

Looking forward to your thoughts!

Give some your thoughts to this discussion(50words)

In all honesty, the concept of Deepfakes really scares me and bothers me. I think generally speaking, deep fakes are very unethical to create, especially when they involve deception, harm or some sort of manipulation with intent to harm. We’ve read and heard of many cases over the years where Deepfakes have been used to harass or defame people.

In terms of protection and safety, I think more regulations and rules have to be made around it. A lot more awareness needs to be raised in the community about what deepfakes are, how they can be created, and what you can do as an individual to protect yourself. As far as schools are concerned, both teachers and students should be made aware of how they can protect themselves from the potential harms of such content.

I think the only time Deepfake could be considered allowed is if there is some sort of consent structure built around it; where there is some benefit to generating deepfakes, conditioned to consent of the people involved.

Question4(150 words)

In the Icebreaker activity, we explored ChatGPT. But there are still many AI tools out there. In fact, https://poe.com is a one-stop site that hosts different AI tools, which include not only ChatGPT but also image generation tools, tex analysis bots, hobby bots, translation bots etc. You are not required to sign up for an account in order to try out these tools.

Other Programs which don’t require an account:  

Perchance

Magic Studio

While personal information such as your name are not necessary for these sites, they do collect cookies and system information such as your IP address.

Alternatively, you can also try out AI educational tools you like, such as MagicSchool.ai (which has a two-week free trial period), Schoolai.com and the AI Image generation in education tools mentioned in this Week’s video, including:

– Ideogram (requires either google or apple sign in)

– Open Art (requires creating an account using email)

– Google Gemini (must sign privacy disclosure agreement, then uses google sign in) 

– Copilot (requires microsoft login- can use UofT email)

– Adobe Firefly (requires account using email) 

– Leonardo (requires apple, google or Microsoft sign in. UofT email applicable)

Other options include Duolingo, or Khan Academy’s AI tutor, all requiring your email to sign up for an account. Or, try to find your own tool! This could be an excellent professional development experience to get to hear about multiple AI systems that we can use in the future. Feel free to check out the systems others are using but try to diversify the conversation so we can all learn about as many as possible.

Here are some prompts to guide your response after you try out at least one AI tool from the bots available in https://poe.com, or the educational AI tools listed above:
(1) What tool did you select to try out? What cool features did you find that is helpful to your studies, professional or daily life? What is the advantage of the features over humans doing the task?

If you use an image generation AI tool, please update the image you generated and prompt you use, and share your experience using it.
(2) Did the tool feel personalized for you? Is there any adpative features? Any errors or downsides for the AI adaptation? How does tool change the way you work or live?
(3) Why does our world need this tool? How do you think this kind of AI tool will change our world in the next 3 years? Or even the next 5-10 years? Any exciting possibilities or concerns you can think of?

The AI tool I checked out was Ideogram, which is very similar to MidJourney. It generates images based on the prompts we enter, but what immediately stood out for me on Ideogram was its ability to interpret text and to write it on the image (MidJourney or Canva most of the times do not write text properly). Tools such as Ideogram can easily be used to generate images, posters, logos and other creative images that can be utilized in the classroom, perhaps to create a certain classroom design, or certain posters to hang in the class.

Give some your thoughts to this discussion(100 words)

I created the two images below using the following prompt: A cartoon illustration of a group of students sitting in a circle listening to their teacher tell a story. The title at the top should say, Morning Circle Time.

This can save a lot of costs in purchasing posters from stores; a personalized poster can be generated and printed at a very low cost and be ready to use in class. I can even see how teachers can use Ideogram to generate manipulatives for teaching certain topics. Now that I think about it, you can even use it to write create your own storybook and have Ideogram generate the illustrations for that story. There’s so much you can do!

What I noticed about the images generated below is that they showed some degree of diversity in the classroom.

Week discussion summary: You will write a response about each topic beginning in Week 3 (250-500 words), summarizing your current thinking (e.g., new insights, ah-ha moments, connections to prior discussions, reflections on your own practices) about AIED. You are expected to make direct reference to concepts/ideas from the readings and notes from the discussion that have influenced your thinking

Guidelines for Overall Evaluation of Learning Journal Entries full Points. Entries consistently show links to ideas from the readings/videos, the course discussion, other students’ reflections and to other online resources if applicable. Entries reflect the learner’s own attempts to grapple with ideas encountered in the course, in relation to the learner’s own development and identity as a graduate student in an education program.