Is AI Hindering Your Learning?

0:00 Look, I know why you're all here.

0:02 You want to know once and for all if using AI in college is good or bad.

0:07 Because depending on who you talk to, AI

0:10 is destroying the age-old institution of higher education.

0:14 Or single-handedly hauling it back out of the grave.

0:18 It's ruining our critical thinking skills.

0:21 Or making us into better writers.

0:23 Or causing an epidemic of plagiarism.

0:27 Or making learning more accessible than ever before.

0:30 In other words, it's confusing.

0:33 But if there's one thing Study Hall's good for, it's

0:35 helping you wade through the confusing parts of college.

0:38 Hi, I'm Erica Brozovsky and here's what you

0:40 need to know about using AI in college.

0:45 [music] Okay, first things first.

0:53 When we talk about using AI in college, there's a lot to talk about.

0:59 We've got educational concerns, ethical concerns, environmental concerns.

1:04 And we'll get to those things.

1:05 But before we do, let's make sure we've got

1:08 a handle on the actual technology and terminology of AI.

1:11 Because AI is a really broad and kind of misleading term.

1:16 It refers to any technology that allows

1:18 computers to do human-like things like learning,

1:21 problem-solving, creating, and making decisions.

1:24 So, AI includes everything from the algorithm that showed

1:27 you this video to the app that predicts your weather,

1:30 to the video generator that made those weird Christmas Coca-Cola ads,

1:34 to the merciless robot overlords that will one day overtake us.

1:38 Kidding about that last one, hopefully.

1:40 What people are mostly worried about when it

1:42 comes to AI in higher education is generative AI, a kind of deep learning model

1:47 that can produce complex content like images, articles,

1:51 sounds, videos, books, and those emails you keep getting from the new HR guy.

1:55 They're worried about it because a lot of people

1:58 are using AI in a lot of ways in college.

2:01 In fact, recent surveys show that around 90%

2:05 of students in the US have used AI for school.

2:08 But exactly how and how much they're using it varies widely.

2:12 Some have used it once, while others are using it weekly.

2:15 Some use it to check their spelling and grammar before turning in assignments,

2:18 and some use it to create study guides or homework plans.

2:21 Some ask for help summarizing long readings

2:23 or searching for information or simplifying complex ideas.

2:27 Some use it to generate research topics and ideas for papers.

2:31 And some are using it to write their papers,

2:33 answer their discussion questions, and even do their math homework.

2:36 It's not just students, either.

2:38 In 2024, about 61% of college professors were

2:41 using AI to do stuff like design curricula,

2:44 come up with new teaching strategies, grade assignments,

2:47 and scan your paper to determine whether you're using AI.

2:51 Whether any of that is actually helpful is up for debate.

2:55 For one thing, AI isn't designed to give you the right answers.

2:59 It's designed to give you answers that sound right.

3:01 See, asking ChatGPT to explain the connection between photosynthesis and Earth's

3:06 atmosphere isn't really the same as asking your college professor.

3:10 It's more like asking your gossipy cousin,

3:13 who heard it from an Instagram post, who heard it from Wikipedia,

3:16 who heard it from National Geographic,

3:18 who heard it from a textbook one time, probably.

3:21 That's because ChatGPT and other models like Gemini, Claude,

3:25 and Grok are what's called large language models or LLMs,

3:29 which means they're really good at generating natural,

3:32 smart-sounding answers to all kinds of different questions.

3:36 LLMs learn to do that by training

3:38 their artificial neural networks on tons of written data,

3:41 from academic papers to Wikipedia pages to fantasy novels to social media posts.

3:47 And they use that training to predict

3:49 how someone might answer the question you asked.

3:51 Lots of times they do give the right answers, but not always.

3:57 They're surprisingly fragile in stuff like math, science, and logic.

4:01 And lots of the writing LLMs train on contain bias like racism and sexism,

4:06 which means their answers can contain that bias, too.

4:09 Ultimately, LLMs can identify how words go together,

4:12 what concepts go with other concepts,

4:14 and what kinds of answers go with what kinds of questions.

4:17 But that's not the same thing as actually knowing the answer.

4:21 LLMs aren't knowledge databases or even search engines.

4:24 They're predictive text generators that spit out words based

4:27 on what statistically is most likely to come next.

4:30 In other words, they're kind of like sophisticated language calculators.

4:33 And so far, the evidence around how AI impacts learning is mixed.

4:39 A range of studies published in 2023 and 2024 have come up

4:42 with a variety of conclusions about the role of AI in learning.

4:45 Like some studies have found that ChatGPT actually makes a pretty good tutor,

4:50 letting you break down concepts, ask follow-up questions,

4:53 and dig deeper into any topic whenever you want.

4:56 Which is great if you're pulling an all-nighter and trying to finish

4:58 assignments when the library is closed and your professor is asleep.

5:01 Other studies show that getting AI feedback

5:03 on your writing might help strengthen your writing skills.

5:06 And then AI might be especially helpful

5:07 for college students with a variety of accessibility needs.

5:10 Some AI models can read text out loud,

5:13 transcribe audio in real time, generate descriptions of images,

5:16 and translate between languages,

5:18 which can make higher education way more approachable for more people.

5:21 There are even AI models designed to help students build

5:23 the social and cognitive skills they need to succeed in college.

5:27 On the other hand, relying too much on AI can

5:31 make you feel more distant from your professors and your classmates.

5:34 And those relationships can be really important for learning.

5:37 Studies also show that college students can use AI as a replacement

5:41 for higher-order cognitive skills like analyzing

5:44 others' ideas and creating your own.

5:46 Using AI to get answers or complete assignments without really understanding

5:50 them can keep you from actually absorbing and retaining the material.

5:53 And it can stop you from building

5:54 important skills like problem-solving and critical thinking.

5:57 So, all of that means that generative AI might be helpful,

6:01 especially for making certain tasks and information more accessible.

6:04 But if you're not actually engaging with the material

6:07 or reflecting on the answers AI is helping you find,

6:10 you might not actually be building helpful skills or learning much at all.

6:14 And that's true no matter what tools you're using to get through courses.

6:18 Okay, here's another way to think of it.

6:20 Say you join a gym and get a trainer.

6:23 But then, instead of learning how to do the exercises,

6:25 you just watch your trainer do them.

6:27 Sure, the exercises are done,

6:29 but you didn't gain any muscle or coordination because you didn't do the work.

6:34 Same goes for academics.

6:36 If you use AI to do your homework, the homework will get done,

6:40 but you haven't necessarily developed your mental

6:42 muscles or built your skills in research,

6:44 making connections, or communicating your ideas.

6:48 If you skip out on the work, you're skipping out on the results of that work.

6:52 And it's not just you who loses out.

6:54 Using AI comes with ethical considerations and consequences that can

6:58 affect you and people in places around the world.

7:01 For one thing, using AI might be considered academic misconduct,

7:06 which depending on your school's policies can mean you fail

7:08 the assignment or the class or even get kicked out of school.

7:12 But using AI can also be really disrespectful to the people who did do the work.

7:17 Now, the legal and institutional lines around who owns AI-generated

7:22 content are super murky and will probably keep on evolving.

7:26 But if you use AI to do work that you pass off as your own,

7:29 you're committing plagiarism.

7:31 See, like you, ChatGPT didn't write that paper all on its own.

7:36 And even if you're not using generative AI to write your whole paper,

7:39 any help you get from AI used data from tons and tons of articles,

7:43 studies, and books that were written by people.

7:47 People who usually have no idea that their work is being used

7:50 to train language models and no way to stop it even if they did.

7:53 LLMs are also notoriously bad at citing their sources.

7:57 So, even if they try not to plagiarize, they'll probably do it anyway.

8:01 Or just totally make stuff up and reference

8:03 books and journals that don't even exist.

8:05 So, along with the consequences to your grade and academic career,

8:09 when you use AI and pass off others' work as your own,

8:13 you risk misinterpreting results,

8:15 obscuring where the information came from, and spreading misinformation.

8:18 To be clear, that's bad.

8:20 On top of all of that are the environmental concerns.

8:23 AI produces electronic waste full of stuff like mercury and lead.

8:28 And the data centers that power AI models use tons of electricity and water.

8:32 Asking ChatGPT a question takes 10 times

8:35 the amount of energy as a standard Google search.

8:38 And according to one estimate,

8:39 AI-related infrastructure around the world could soon consume four to six

8:44 times the amount of water as the entire country of Denmark.

8:47 Meanwhile, we live in a world where

8:49 a quarter of humanity lacks access to clean water.

8:52 All those ethical quandaries have made lots

8:55 of people eager to get some regulations on AI.

8:57 And colleges are no exception.

9:00 Higher education has reached a huge moment of reckoning with AI.

9:03 Like you, they're asking a lot of questions about using AI in college.

9:08 Like should students be allowed to use it?

9:10 When?

9:11 How much?

9:12 Could they even stop you if they tried?

9:14 And so far, they haven't really come up with much consensus.

9:17 Usually, policies about AI use are left up to professors.

9:21 So, even different classes at the same school might have different policies.

9:24 Some say you can use generative AI however you want.

9:28 Some say you can't use it at all.

9:30 Most seem to fall somewhere in the middle,

9:32 saying you can use it with some considerations and restrictions

9:35 to try to keep it helpful and ethical and, you know,

9:38 avoid possibly failing the course.

9:40 Which brings us back to where we started, confusion.

9:44 Unfortunately, it's hard to say once and for all if

9:47 using AI is good or bad or, I don't know, some secret third thing.

9:52 AI has some really interesting potential benefits, but it also comes with risks.

9:57 Not to mention, if you're using ChatGPT to write

10:00 your paper and your professor's using ChatGPT to grade it,

10:03 well, then what are we even doing here?

10:05 AI is complex and it's changing by the day.

10:09 So, we don't even really have the whole picture yet of what it means

10:12 for you or the future of higher education or the state of the world.

10:16 As AI continues to develop,

10:19 the impacts and costs might change and so will the way we think about them.

10:23 So, for now, the question for college students, professors,

10:26 and really everyone to wrestle with isn't just is AI good or bad?

10:31 Instead, there's a range of questions we can ask now and in the future

10:35 to help us figure out when and how to use AI.

10:38 What do I think college is for?

10:39 And what do I want to get out of my education?

10:42 Is using AI building my thinking or replacing it?

10:45 What do I value?

10:46 And how do I want to interact with technology in the world around me?

10:49 And finally, how does using AI right now affect me,

10:53 other people, or the planet in the future?

10:55 Because ultimately, what we do now and what we choose to learn

10:59 is part of that question we've been wrestling with pretty much forever.

11:03 What does it mean to be human?

11:05 If you're currently in or thinking about going to college,

11:09 go to gostudyhall.com for more student resources

11:11 and check out our $25 college courses.

11:13 Thanks for watching.

11:14 See you next time.