AI-Generated Video: Using Technology in Authentic Assessment

Transcript:

Good day!

In this short session today, you will learn about using technology in authentic assessment.

Our goal is to achieve three things.

First, you will be able to distinguish a valid online assessment that ensures data is processed through the brain; from an online assessment that is invalid.

Second, you will be able to demonstrate the proper use and understanding of AI by writing at least one concise prompt to automate work.

And third, you will be able to explain the importance of technology in reducing non-instructional or non-mentoring tasks.

For this work, we will need the following materials.

First, we need test papers from students, the test papers have to be clearly photographed or scanned and are ready for uploading.

Second, we definitely need a document containing rubrics for the assessment that we have just finished.

And last but not least, we will need generative artificial intelligence—a multimodal artificial intelligence. For today, we will use Google Gemini as our choice of AI.

Now that we have completed our list, let's begin with the first material for today's activity.

Here, we have a test paper after completing an assessment with students. Take a moment to observe the test paper.

As you may notice, there is no name given, only a student code written at the top to help eliminate biases . The name of the assessment is written on the second row of the paper.

Another thing to observe is the answers. Here is one answer, and here is another answer.

To help you describe it, let's consider how the answers were submitted. Were they encoded on Moodle and submitted through the LMS? Or was the assessment a digital copy of the student's work created on a word processor or another online document editor?

Or is it the the student submit a printed copy of a document from their computer?

If you answered none of the above, then you are correct. None of these options describe these test papers.

These papers were actually handwritten by the student. The learner may have written it as a result of their own thought processes, or they may have copied from something they saw, heard, read, or even from someone else. But one thing is certain: it went through the student's mind.

Learning was affected, and what appeared on the paper is exactly what was present in the person's mind. Let us pause this video after the question for reflection: Where could have the answers come from if instead, the students wrote it straight in the LMS, or submitted an electronic document or its printed copy? Done answering? Now, that spells the difference between valid and authentic assessment versus an invalid ones.

With that, let's continue.

Moving on to the second material, we also have a document that contains the rubrics for the assessment that was given.

We need to evaluate and grade the answers using well-written rubrics.

For this assessment, our rubrics are as shown. We assign three points for effort, three points for logical coherence, and the most important part is reflective learning and absorption of the lesson, which should have the highest number of points, . We also include, below the article, as reference for the answers provided by the learner. Before we proceed, think about it: would grading the papers be easy without this rubrics? You may pause the video to answer the question, as well as to write your own rubrics on your favorite word processor. Please save it as a PDF file.

Now, let’s proceed to the third important material we will use, which is our AI. Since we are using Gemini, please open your browsers now, go to Gemini by typing gemini.google.com/app, and click on “new chat” to start prompting.

Our prompt should be concise and clear so that the AI can understand and execute it properly. This process is called prompt engineering. We begin by specifying the role the AI should take.

In this case, we will type: “You are a teaching assistant who specializes in checking essays for an instructor.”

Once the role of the AI is established, we need to give the AI instructions on what to do. The instruction goes like this: “Attached are the rubrics for checking test papers.”

We will use the rubrics that we have just created right here. There it goes. Let's save it.

Now, we upload it to Google Gemini.

Now that it's been uploaded, let's proceed with the next instruction.

Grade these student papers based on the provided rubrics. Once confirmed, we move to the next instruction.

We need to be concise. For each paper, extract the student ID found on the first row, and extract the name of the assessment found on the second row of the test paper.

Then, calculate the total score of the quiz and format the final output as a table with two columns.

So, what would the columns be? The student ID and the actual name of the assessment. The scores will be placed in the second column.

Now that we have completed the prompt that will be used by AI, let's submit it.

Wait for the analysis. While waiting, you may pause this video and follow the same prompt on your AI, and try to observe the characteristics of a good prompt that will help you automate your work.

Have you done it? Great!

In here, the AI has received and reviewed the rubric for the assessment. The ones we uploaded are ready for grading to ensure the highest accuracy. We have broken down the criteria as follows.

Effort: maximum of three points. Three points for more than one sentence, two points for one sentence or less.

Logical coherence: three points for well-explained answers. Two points for answers that provide reasons but are unconvincing or lack logical support.

For reflective learning: the scores are four, two and one, based on its similarity to the key article provided. .

So, the total score potential stands as outlined.

Now it’s time to upload the papers. Let’s add files, upload the files, and get those test papers. One, and two. Done.

Then, click open.

We have uploaded those test papers. And submit.

Here we go, let's hope for the best.

B one has three points, three points, and four points. A perfect score. B two had three points, three points, and one point, and the reasons are stated as well.

Here is the table which we can export to Google Sheets.

So, instead of checking them one by one, we can now upload all those test papers after taking a picture and export them to Google Sheets. as well. This allows us to integrate them with our digital class record.

Now, let's consider the question: how much time would be saved by using this method compared to reading those essays manually? What can you do with the time saved?

On the other hand, if we don't read those essays, and if we don't let students write essays by hand, they might just copy and paste someone else's work. This means the answers may not reflect their own thoughts.

Therefore, this authentic assessment can be digitized and expedited using the right technology. Right now, that technology is called AI.

Were you able to follow the steps we just shared? If not, try again, or do it now. You can pause this video and continue as soon as you're done.

Have you done it? Very good.

To recall, you have just done the following:

You were able to distinguish a valid, authentic assessment from a nonvalid ones.

You were also able to try the proper techniques in AI prompting, and understood how it works.

And finally, I can leave you with this question: Can you now explain how this technology is important to your colleagues?

if so, please do it so that we can make this world a better place,

let's not stop here.

Let us share good news

and have a great day.