The West Chester University Teaching & Learning Center has created a set of examples to test your AI detection abilities.
Listen to both audio samples taken from the Strike Force Five podcast and determine which one is AI and which one is real.
Submit your guess using the form below and find out how you did.
The following text was submitted to the Turnitin AI detector which indicated that it was 80% AI written.
3.1 Uniform Crime Report
This chapter will teach us how to create a new variable and produce summary statistics. We will use the crime data from the Uniform Crime Report (UCR)—specifically, the 2018 Part 1 crime data from Pennsylvania. More than 18,000 police departments in the US report crime data to the FBI, and the FBI compiles the nationwide data and publishes the UCR.
Crime analysts need to know the differences between Part 1 crime and Part 2 crime. There are two categories of criminal offenses. Part 1 offenses, known as index crimes or serious crimes, are generally felonies that can result in more than a year of incarceration in prisons. Some violent and property crimes that fall under Part 1 offenses include homicide, rape, robbery, aggravated assault, burglary, larceny-theft, motor vehicle theft, and arson. On the other hand, Part 2 offenses are non-index crimes and are considered less serious. These crimes are generally misdemeanors that warrant less than a year of incarceration. Part 2 offenses may include simple assault, vandalism, fraud, drug offenses, disorderly conduct, and other misdemeanors. The biggest limitation of the UCR is that the data only counts crimes reported to the police, so you cannot know how many crimes are underreported. This unreported crime is known as dark or hidden figures of crime, and several data-collecting strategies have been developed to estimate this underreporting. I will introduce these datasets in different chapters. In this chapter, we will learn how to compute descriptive statistics using Part 1 crime data.
Is the Turnitin AI Detector correct?
You are incorrect.
This text was 100% human written by someone who does not speak English as their first language. This is an example of how AI detectors are more likely to misidentify human writing as AI generated from individuals whose first language is not English OR who come from homes where English is not the first language spoken. The TLC has encountered other examples of this problem with our WCU students.
You are CORRECT!
This text was 100% human written by a WCU faculty member who does not speak English as their first language. This is an example of how AI detectors are more likely to misidentify human writing as AI generated from individuals whose first language is not English OR who come from homes where English is not the first language spoken. The TLC has encountered other examples of this problem with our WCU students.
Here are two examples of responses to an assignment prompt used in our Online Faculty Development Program. Both had zero percent AI written when submitted to TurnitIn's AI detector. One is real and the other is generated by AI. Can you tell which is which?
Looking back at my time as a student, one moment that sticks out is from a massive history lecture course I took in my sophomore year of college. This class had over 150 students, and the instructor, Dr. Thompson, was all about calling on the same few front-row students who always seemed super confident. Meanwhile, I was one of the quiet ones, hanging out toward the back. There was this one discussion about the Civil Rights Movement where I really wanted to share my thoughts because I had done some volunteer work related to the topic and felt like I had something unique to add. But nope, Dr. Thompson never called on me, even though I had my hand up for what felt like an eternity.
I left that class feeling pretty bummed out, like my voice didn't matter. If I could go back and give Dr. Thompson some advice on being more inclusive, here's what I'd say. First off, mix up how you get students to participate. Maybe throw in some online discussion boards where everyone can post their thoughts at their own pace. This way, quieter students like me or those who need a little more time to think things through can still get their say.
Second, work on creating a more welcoming vibe in the classroom. Start with some icebreakers or small group discussions to help students get to know each other. In a big lecture setting, breaking into smaller groups once in a while could make a huge difference.
Third, use a random selection method to call on students instead of just relying on the same volunteers. This makes sure everyone gets a shot to participate and helps quieter students feel included.
Lastly, mix up your assessment methods. Let students show what they know through different formats like projects, presentations, or written reflections, not just traditional exams. This caters to different learning styles and strengths.
By doing all this, instructors can create a more inclusive and supportive learning environment. It's not just about hearing from the loudest voices but making sure every student feels valued and heard. This approach doesn't just improve individual experiences but also makes the whole class better.
It took me a minute to think of a time when I felt excluded. I am privileged to have felt supported in most academic settings. I always did well in school, I'm not neurodivergent, and I'm a white, cis-het woman. I am very good at doing school.
However, as I thought about it, I recalled a time in elementary school when I was mercilessly teased by my peers. I'm not sure why they singled me out -- I didn't stand out from my peers in any notable way. However, it was horrible. I recall one day one of the boys said something crude that made me feel really uncomfortable (though if you had to ask me, I wouldn't remember what he said or why it hit differently). I took it to the teacher, who said that boys will be boys and that I should get a thicker skin. Now, this was in the early '90s. Today, the boys will be boys line is much less acceptable. But it made me feel so isolated and unsupported. My feelings didn't matter, even if they made me feel like the classroom was not a safe place to be.
If I could go back, I would tell the instructor that all students need to feel safe in the classroom. This includes feeling safe from teasing and inappropriate remarks. I would advise her to acknowledge my feelings as valid. I would then suggest that she have a private meeting with the male student to explain why their behavior was unacceptable. I would also suggest that she have a larger class policy that these behaviors will not be tolerated. This should be part of a larger discussion at the start of the school year where the teacher explains the rules of the classroom.
I know in my own classroom, I establish from the beginning that sexist, racist, transphobic, and other such remarks are not tolerated. I am clear with students that if they make the classroom unsafe, I will escalate their behavior to the appropriate offices (I'm not sure which office yet at WCU!) I have not yet had to do this, but I take making a safe learning environment seriously.
Submit your guess using the form below and find out how you did.
I told the AI to change what was written to be in the style of Eagles beat reporter Jimmy Kempski. That simple prompt, to change the writing to be in a specific style of a person is often enough to drive the AI detection down to zero.
For individualized support, reach out to your assigned instructional designer or email us at TLC@wcupa.edu.
To engage with the Teaching & Learning Center's AI content visit the TLC blog or the TLC website.