The Rise of AI - Implications to Current Available Courses and Future Jobs | Chesa Caparas (Part 1)

Summary

In a recent talk by Chesa Caparas, a faculty member at De Anza College, the implications of artificial intelligence (AI) on education and the workforce were explored. The talk, aimed at providing a general background on AI, delved into its impact on educational practices, career transformations, and the ethical challenges it poses. Caparas highlighted the significance of AI literacy in modern education and drew attention to the potential restructuring of industries like healthcare due to AI advancements. The necessity for critical evaluation of AI outputs and the importance of human connections in the digital age were emphasized.

Highlights

• Chesa Caparas discusses AI's broad implications on modern education and job sectors. 🎓
• Generative AI can produce human-like content, affecting jobs traditionally done by humans. ✍️
• AI's use in healthcare promises improved diagnostics but raises privacy and ethical concerns. 🏥
• AI outputs can be unreliable and biased, necessitating critical evaluation from users. 🤔
• Emphasizing human qualities and connections is essential amid growing AI reliance. ❤️

Key Takeaways

• AI is transforming education and the workforce, requiring new literacy skills. 📚
• Generative AI, like ChatGPT, creates new content and impacts various professions. 🤖
• AI's influence extends to industries such as healthcare, altering diagnosis and treatment processes. 🏥
• Privacy concerns and the trust erosion in AI outputs demand critical evaluation skills. 🔍
• Human intelligence and critical thinking remain crucial in navigating AI-driven changes. 🧠

Overview

Chesa Caparas, in her insightful presentation, outlines the rapidly evolving role of artificial intelligence in education and beyond. She explains how AI, particularly generative forms like ChatGPT, is producing content that once relied solely on human effort. This technology is not only changing how students learn but also how educators and professionals across industries adapt to new realities.

Particularly intriguing is the discussion of AI's encroachment into healthcare. Caparas illustrates how AI can now assist in diagnosing diseases and tailoring treatments based on data patterns. However, she also highlights significant concerns about patient privacy and the potential loss of human touch in patient care, emphasizing the need for an ethical approach to AI integration.

Throughout the talk, the importance of viewing education as a space for developing critical AI literacy is underlined. Caparas advocates for strengthening human connections and the unique problem-solving abilities that humans possess, which remain indispensable despite technological advancements. She encourages embracing curiosity and effective communication as a bridge towards becoming more adept in an AI-driven society.

Chapters

• 00:30 - 01:00: Introduction and Speaker's Background The chapter introduces the speaker, Jessica Paras, who is a faculty member at De Anza College specializing in English, Asian American studies, and women's studies. Additionally, she mentions her experience as a Fulbright scholar where she conducted research on media and information.
• 01:00 - 01:30: Research Interests and Talk Overview The chapter focuses on the speaker's research interests and the overview of their talk. The speaker is interested in digital literacy, specifically teaching students to effectively and ethically use new digital tools to improve their own lives and the lives of others. The talk covers AI education and the future of work, beginning with a general background on artificial intelligence.
• 01:30 - 02:30: Impact of AI on Education and Careers The chapter titled 'Impact of AI on Education and Careers' explores the profound effects that artificial intelligence (AI) is having on both educational environments and career landscapes. Initially, the chapter delves into how AI is reshaping student activities and the broader educational culture, questioning the traditional purposes and methods of education in the face of technological advancements. Furthermore, it analyses the shifting dynamics in careers, examining how AI innovations are altering job roles, responsibilities, and potentially redefining entire industries. The discussion underscores the urgency of understanding AI's implications, as emphasized by a poignant quote highlighting the stakes involved.
• 02:30 - 05:00: Understanding AI and Machine Learning The chapter discusses the impact of modern technologies, particularly AI and machine learning, on human consciousness and behavior. It emphasizes the idea that the technologies we use can shape our minds, as suggested by Johan Hari in his book 'Stolen Focus.' The text underlines the importance of understanding these technologies as they are fundamentally altering not only our thought processes but also our actions. The chapter encourages the reader to reflect on life before the advent of technologies like the cell phone to appreciate the magnitude of these changes.
• 05:00 - 09:00: Introduction to Generative AI and Large Language Models The chapter introduces the concept of Generative AI and Large Language Models, emphasizing their transformative impact on thinking, behavior, and perceived possibilities. The speaker points out that while AI, often referred to as the fourth Industrial Revolution, is frequently discussed in the media, the discourse is often vague and imprecise.
• 09:00 - 12:50: Impact of AI on Jobs The chapter titled 'Impact of AI on Jobs' begins with addressing common fears and misconceptions about AI, notably the dramatic scenarios where AI could destroy humanity or cause massive job losses, akin to scenes from the 'Terminator' movies. However, the speaker clarifies that such fears are exaggerated and not currently realistic. The chapter aims to define AI clearly to provide a better understanding of its current capabilities and limitations in the context of job impact.
• 12:50 - 20:00: AI in Healthcare The chapter titled "AI in Healthcare" discusses the origins of the term "artificial intelligence," which was first coined by John McCarthy in 1955. McCarthy defined AI as the science and engineering of making intelligent machines, although the definition can be somewhat vague. The chapter seeks to delve deeper into understanding what constitutes an intelligent machine and its implications, particularly in the context of healthcare technologies.
• 20:00 - 23:00: Structural Changes and Privacy Concerns The chapter begins by defining intelligence according to the Oxford Dictionary as 'the ability to acquire and apply knowledge and skills.' This sets the foundation for discussing machine intelligence, which is similarly described as a machine that can learn and apply knowledge in new situations. This concept is crucial in understanding the structural changes and privacy concerns associated with the integration of machine intelligence into various sectors of society. The chapter likely elaborates on how these changes impact privacy and what implications they have for both individuals and institutions.
• 23:00 - 29:30: Trust and Ethical Concerns in Education The chapter titled 'Trust and Ethical Concerns in Education' delves into the realm of artificial intelligence, focusing on a significant subset known as machine learning. It emphasizes the training of machines to acquire new skills, aiming to enhance their perception, knowledge, decision-making, or actions based on accumulated experience or data. The chapter underscores the reliance on pre-existing knowledge while incorporating data-driven experiences to refine AI capabilities, suggesting implications for trust and ethical considerations within educational frameworks where AI is increasingly integrated.
• 29:30 - 37:30: Understanding AI Limitations and Bias The chapter titled "Understanding AI Limitations and Bias" explores the basics of artificial intelligence, starting with a general definition that mirrors human learning processes. It highlights how AI systems, similar to humans, learn, perceive, decide, and act on knowledge. The chapter seems to set the stage for a deeper dive into specific types of AI, notably generative AI, though the provided transcript ends before detailing this subset's features or its relation to AI limitations and biases.
• 37:30 - 46:30: AI Literacy and Education The chapter titled 'AI Literacy and Education' introduces the concept of generative AI, focusing on platforms like ChatGPT. It discusses the role of AI systems in creating digital content across various forms such as language, images, video, and music. The key feature of generative AI highlighted is its ability to produce new and original content autonomously, differentiating it from traditional AI that may only process or analyze existing data.
• 46:30 - 51:00: Human Intelligence and Critical Thinking The chapter discusses the traditional role of human intelligence and creativity in generating text and images, and contrasts this with the capabilities of generative AI. It explains that tasks humans used to perform, such as writing and image creation, can now be accomplished by AI through the use of large language models. The chapter aims to provide an understanding of how these models function.
• 51:00 - 54:00: Conclusion and Encouragement of Curiosity In this chapter, the concept of pattern recognition is discussed. A sequence of shapes in various colors is presented with a missing final shape. The pattern follows a recurring order: yellow square, red triangle, blue circle, and so on. The reader is encouraged to identify the next shape in the sequence, which logically should be a blue circle. This exercise is used to illustrate the importance of observing patterns and encourages curiosity by engaging with interactive problem-solving.

The Rise of AI - Implications to Current Available Courses and Future Jobs | Chesa Caparas (Part 1) Transcription

• 00:00 - 00:30 [Music] foreign [Music] good afternoon everyone my name is Jessica Paras and I am faculty in English Asian American studies and women's studies at De Anza College I was also a Fulbright scholar through feu and I researched media and information
• 00:30 - 01:00 literacy last year so my research interests include digital literacy or how to teach students to better utilize new digital tools how to use them ethically and effectively and to improve their livelihoods and the livelihoods of others so my talk today is on AI education and the future of work and the goals for today are to First just give a sort of General background of artificial
• 01:00 - 01:30 intelligence I think a lot of you have probably heard about it but also talk about the way it's impacting education in terms of what students are doing and what it means for the culture the purpose of education and also talk about how careers are changing as a result of artificial intelligence because there's a lot happening so but I want to start off with this quote because I think it'll show us what the stakes are why we have to understand what artificial intelligence is and how
• 01:30 - 02:00 it works so Johan Hari in his book stolen Focus says take care what technologies you use because your Consciousness will over time come to be shaped like those Technologies so it's important to understand how these new technologies work because they are changing our brains which means they're changing not just how we think but also then how we act in the world and um if you know any of you have remember what life was like before a cell phone
• 02:00 - 02:30 maybe maybe I'm just old then you'll see that that those new technologies have really transformed not just how we think but also how we act and what we believe is possible so that's why I want to talk about AI today and you've probably heard a lot about AI in the news people are talking about it as the fourth Industrial Revolution but what's sort of I've noticed in these reportings is that it's kind of an imprecise way of talking about AI
• 02:30 - 03:00 there's a lot of like kind of oai is coming to destroy Humanity or it's going to take all our jobs or you know we're going to have Terminator running through the streets uh that's not exactly accurate um yes there are you know new robots being trained and you know new forms of machine intelligence but we're not at Terminator stage yet so I I want to First give some definitions of AI to help us understand kind of what are
• 03:00 - 03:30 these technologies that we're working with John McCarthy was one of the co-authors of the paper that first used the term artificial intelligence this was back in 1955 okay and he called it the science and engineering of making intelligent machines okay which is kind of helpful but also not very helpful because what does it mean to have an intelligent machine well it's probably more helpful if we also understand what is
• 03:30 - 04:00 intelligence so according to the Oxford Dictionary intelligence is the ability to acquire and apply Knowledge and Skills so you know as students we become intelligent as we learn new things and apply them in our lives same thing with machine intelligence right it's a machine that can learn things and then apply that knowledge in new situations so another
• 04:00 - 04:30 term you might have heard in relation to artificial intelligence is machine learning okay and this is where we're really thinking about how machines are being trained to learn new skills and the definition is it's the part of AI that studies how computer systems can improve their perception knowledge decisions or actions based on experience or data okay so you can see that a lot of the definitions are really drawing from what we already know about
• 04:30 - 05:00 intelligence how do we learn new things right we get new information we perceive new things we make decisions and we act on that knowledge so that's a very general definition of AI or artificial intelligence what most people are talking about now is a smaller form or like a subset of artificial intelligence and that is generative AI that's another form of AI and
• 05:00 - 05:30 probably the most popular form right now is if you recognize this logo chat GPT maybe some of you have it pulled up on your phone right now but generative AI refers to computer systems that can produce or generate various forms of traditionally human expression in the form of digital content including language images video and music so what makes it different is that it generates new content okay
• 05:30 - 06:00 and so it's really important because the the definition that says it's like traditionally human expression these are things that humans used to be able to generate we used to have to depend on humans to write new text right to create a new image but now generative AI can do it and the way that it does it is through what's called large language models so maybe some of you have heard of the term large language model but I want to show you how they work by giving you a little
• 06:00 - 06:30 test okay um fill in the blank you see we have a sequence of shapes in different colors here and there's a blank what should go in the last slot there what should go in that blank Circle what color blue okay how did you know it's patterns okay so we notice a pattern we see yellow Square red triangle blue circle yellow Square red triangle so blue circle must come next
• 06:30 - 07:00 okay so what you did it's similar it's a very simplified version but it's very similar to what the large language models do which is they look for patterns and they try to predict what comes next okay and what you what when you mentioned how you figured it out it's because you were using certain parameters meaning you were thinking about okay what is the shape and also what is the color and what is the sequence you know what comes after the other okay so think of a large language model as a similar machine that uses
• 07:00 - 07:30 parameters except instead of shape and color it uses parameters like how often a word appears in a sentence or appears in a paragraph how close it is to other words how often it appears with different words in the same chapter so those are how it makes its predictions it uses like millions of parameters lots and lots of parameters to figure out what is the appropriate text or that
• 07:30 - 08:00 should come next okay it's very exciting but these large language models have been adapting so quickly that they are starting to impact a lot of our jobs right think about if we used to have humans writing text and creating images now that we have machines that can do it it is going to impact the work that we do okay so there's one study of the impact on jobs in in the U.S that says AI models including gpt4 which is
• 08:00 - 08:30 was until maybe like two days ago the most recent version of GPT um all of those software tools would heavily affect 19 of jobs with at least 50 percent of the tasks in those jobs exposed right so what that means is 50 of the work that you would do is now exposed to the changes in Ai and some of that is great some of it is like oh great I don't have to write a million emails anymore chat gbt you saved my life okay but
• 08:30 - 09:00 what's interesting about these new Revolutions in technology is that in contrast to what we saw in earlier waves of automation higher income jobs would be the most affected so we used to have you know sort of more manual labor being automated you know think about cars and you know different uh things that we would construct that have been automated with machines with large language models and things like
• 09:00 - 09:30 chat GPT white collar or higher income jobs like paralegals writers content creators all of those are the ones that are more exposed to the changes with these new technologies even software Engineers they're basically writing themselves out of a job but there you go but it doesn't have to be completely destructive right if you learn how to use the tools if you learn how to work with them they can actually make your job easier better you can they
• 09:30 - 10:00 can make you more efficient the real issue is just knowing how to use them ethically and effectively so I want to give you an example of one of the sort of industries that is being impacted and I think you know in a Philippine context it's probably a really important industry to talk about but Healthcare right a lot of new technologies are being adopted in healthcare probably one of the easiest ones to
• 10:00 - 10:30 predict is like transcribing patient notes because chat GPT can write text really easily and really quickly it's like I don't want to have to write every patient's story what it will do is it will write a sort of template okay and then you just have to kind of add in different uh specifics for that patient that you're working with okay but what's really interesting about Ai and Healthcare is also the ways that it's being used for detection diagnosis and treatment so remember that
• 10:30 - 11:00 these are pattern making machines like they try to spot patterns and then fill in the blank so in terms of detecting new diseases or you know let's say cancerous cells right it uses these predictive models right and same thing with diagnosis as opposed to just giving a patient a generic oh you have diabetes let me give you you know the generic treatment for diabetes it uses more statistical data and modeling to say
• 11:00 - 11:30 well based on your income your gender where you live and you know your family history it's most likely that this type of diagnosis and treatment is the right one for you okay so it's allowing us to have a little bit more of a specific um response to patients in the healthcare setting and more tailored to their specific experience however remember
• 11:30 - 12:00 that it's also based on previous data patterns that it's seen before and I'm going to talk a little bit about what happens when you don't have accurate data to make your predictions but the one thing I do want to emphasize with this example is thinking about AI not just as a new tool but as a structural driver of change so what that means is it's not like having a nicer pencil right but it's like imagine the
• 12:00 - 12:30 changes that came around with the phone right or with the car it didn't just allow you to do what you did a little bit faster it changed the whole social structure right think about a car for example it's not just like oh well now I can get there faster than walking but it literally changed our structure and that we could start to build things further away we could start to travel further away we had to build new infrastructure so that is kind of the changes that are going to come with
• 12:30 - 13:00 these new AI tools the very structure of healthcare could change but there are problems there are definitely limitations okay and you've maybe heard some of these concerns as well if you've been reading the news the first is privacy and this idea of the ideal patient because Ai and these models depend on lots and lots of data they then need lots of your data so all
• 13:00 - 13:30 of your health history all of your family's health history any concerns that you might have had any like weird test results all of that they want all of it and so what's happening is that now these companies want more and more of our information and especially like our biometric information if it's in the context of Health Care and that can have really big concerns for privacy right because maybe you don't want a machine to know everything about your personal medical history and then make
• 13:30 - 14:00 predictions about it because we have seen that that can severely limit the access to health care that some people have you know if you think about pre-existing conditions another concern is the inability to validate results from generative Ai and if any of you have ever used chat gbt to write an essay I'm sure you haven't because you're all you know really honest students you do your own work but if you know someone who has has used chat GPT you'll know that it doesn't produce the
• 14:00 - 14:30 same thing every time right even if you ask it the exact same question it will give you a different answer a different essay so similarly when it's producing diagnoses or treatment plans it's really hard to validate whether this is the best one because each time the model spits out something different so you're like is this the one that I should be using how do I know that this is the ideal one and of course the problem is chat gbt and other large language models can't explain their own thought process
• 14:30 - 15:00 you know that's one of the things that keeps us smarter than them and of course another big concern and for me this is the biggest one is the loss of that human connection right because when we think about health care we think about how it's really about building relationships you know between doctors nurses patients you know Healthcare Providers and those who are seeking care and when it we're talking about care we really want to talk to a human right we want to feel like a human sees us knows
• 15:00 - 15:30 us understands us not just a bot that says as a large language model I cannot answer that question for you you know you want something real you want to feel like there's that connection and so that's also what I want to emphasize in this presentation as we become more and more I don't want to say dependent on artificial intelligence but as it becomes more embedded in everything we do we need to really
• 15:30 - 16:00 over like emphasize our human qualities and our human connections because there are larger ethical questions when using these AI tools right and that is the erosion of trust any of you have been in a classroom in the last year you have probably heard teachers say oh my gosh I can't believe the student needs chat gbt or I think the student is cheating I think the student is using Ai and I you
• 16:00 - 16:30 know they're not learning anything they don't care about writing anymore there is really this culture of um and I what I've witnessed in my own institutions is educators really questioning the Integrity of the students okay just just because they know how easy it is to use chat GPT and they think well you know why wouldn't they use it okay um but that leads to a larger problem as
• 16:30 - 17:00 well with students also not trusting uh the relationship with the teacher so I have a lot of students who say one of their biggest fears is being accused of cheating you know so there's another level of stress or anxiety within the students because they're like even if I do really hard work if I write a good essay the teacher will just think that I've cheated so I think that's important to understand that we still need to find a
• 17:00 - 17:30 way to have that human connection to keep that trust because another thing that's going to happen if students depend more and more on using these AI tools they're going to start to question the purpose of Education right why should I go to school when I can just learn everything or chat GPT can write everything for me and summarize everything for me so we have to maintain trust in the education process one thing that we definitely should mention with
• 17:30 - 18:00 artificial intelligence is how it can hallucinate so that's another aspect of trust and I'm not sure if folks have heard that term before like how chat gbt or AI can hallucinate basically it can make up false information know it's not grounded in truth since it's grounded in probabilities it just gives you an answer that's you know more statistically probable right it's not necessarily true if you ask it for
• 18:00 - 18:30 um you know a recommendation for a recipe it'll give you not something that it knows tastes good it'll just give you something that you know it has matched a pattern from previous recipes and tries to find that oftentimes okay but sometimes you cook it and you're like I don't know what this is I don't know what they were thinking right so it can hallucinate it can create falsehoods really important to remember it is not grounded in fact
• 18:30 - 19:00 and it can also be specifically intentionally used to create falsehoods right A lot of people have been using AI to create deep fakes because you can use artificial intelligence to record someone's voice you just need three seconds of a recording of their voice and then AI can create an entire conversation using that voice so there have been scams that people have done
• 19:00 - 19:30 where you know let's say you call someone up on the phone even if it's the wrong number and just the three seconds where you say hello sorry you have the wrong number AI can then build an entire conversation using your voice from those few seconds and they've used that to then you know create a a scam where the scammer will call your parents and pretend to be you and say oh I've been kidnapped and I need money send me money okay that
• 19:30 - 20:00 you've also heard of deep fakes yes maybe you've seen some deep fakes right so AI can do that as well all it needs is just like a small recording of your face your facial expressions you talking and then it can generate an entire video on that and it can also use your voice so that has a lot of implications for how we move through the world right again this is a structural driver of change it's changing the structure of our world and I think you know again
• 20:00 - 20:30 folks here in the Philippines are no strangers to the challenges of Miss and disinformation okay so um Martin Ford in his book rule of the robots he's talking about deep fakes and he has this great statement saying we will have to somehow learn to navigate within a new and unprecedented reality where what we see and what we hear can always potentially be an illusion okay so
• 20:30 - 21:00 it's kind of stressful maybe a little scary to think that the world that we're moving into is one where potentially everything is an illusion or false right a fabrication of reality using artificial intelligence right you might I think that really looks like Chessa maybe that's her but maybe someone will take this recording and then create some you know deep fake of me saying that you should I don't know do something really
• 21:00 - 21:30 bad don't believe it if it's telling you to do something bad it's not me okay stay in school do your you know all that good stuff so so anyway um it's again this is part of that erosion of trust right when we have a culture that is surrounded by we don't know what's true anymore it can be really stressful but I don't want to be like all doom and gloom right because that doesn't mean that we have to be walk we don't have to walk through the world afraid what we really need to do is understand how
• 21:30 - 22:00 these falsehoods are generated and how they work okay so let's go back to that idea of the large language model and how it's being trained um because when you have a model that's trained to find patterns if there's a piece of data missing then of course you're going to get falsehoods right if you're trying to predict diabetes in a population but you only have data from a specific race okay then your patterns
• 22:00 - 22:30 are not going to be applicable to people of A different race okay and that's one of the things that is has actually been concern with artificial intelligence so remember um this pattern that we saw and what was the answer again blue circle right how many sure yeah definitely blue circle but what if we think that this is not a pat a sequence of three things being repeated what if we saw it as a sequence of six
• 22:30 - 23:00 things being repeated you know what if the sequence is square triangle circle square triangle star and then that is what gets repeated okay so we wouldn't know that because we didn't see we didn't have enough information we didn't see the rest of the sequence we made a prediction based on what we saw but we didn't see everything and that is also one of the concerns with AI right especially generative AI which is trained on stuff that's on the internet and I don't know if you've read the
• 23:00 - 23:30 internet lately but there's a lot of bad information out there okay if we think about it that way when it's missing data you're not going to get an accurate model you're not going to get an accurate prediction and there are a lot of cases where this has happened where tools that have been trained on sort of bad data or missing data are then producing something that's you know perpetuating that misinformation and so
• 23:30 - 24:00 I want to give you an example in 2013 un women ran a campaign called the autocomplete truth of folks familiar with autocomplete you've seen you type in a Google Search and it just fills in the rest of it for you like oh my gosh how did it know that that's what I wanted to ask it right um but the campaign was actually to see how Google autocomplete thought about women if you put in the phrase women should or
• 24:00 - 24:30 women need to it'll auto complete for you and so some of the results they found is when you put in women need to it says women need to be put in their place women need to know their place women need to be controlled or if you typed in women shouldn't it auto-completed women shouldn't have rights women shouldn't vote and then same with women should women should stay at home women should be
• 24:30 - 25:00 slaves so not none of these are true okay but they are statistical probabilities that Google autocomplete found it's really important to understand you know I think a lot of times when we go to Google we sort of assume that it's going to give us something close to what we're looking for right but if we stop to think about how do these models actually work then we can think critically about why did it
• 25:00 - 25:30 complete that you know why did it complete that sentence in that way and we'll say that oh well a lot of the stuff written on the Internet is written by Men actually the majority of content on the Internet is written by men so that is why you might see things that have maybe a more misogynistic view to them when you get the autocomplete results so I give this example not just to say that you know AI is bad and it has a lot of faulty information it gives
• 25:30 - 26:00 falsehoods I give this example because I also want to say that when people organize and draw attention to these limitations of the technology things can change so after the UN ran this campaign Google revised its search so that it would no longer reproduce misogynistic content right so it would you can actually curate the AI you can actually dictate you know whether or not it's
• 26:00 - 26:30 gonna reproduce misogynistic or racist or any kind of hateful content so there is the possibility of Regulation so knowing that if we know how the technology Works we're aware of its limitations and we're able to draw attention to the problems within the technology we can make changes that are less harmful to people and that is what I think the role of education is in this
• 26:30 - 27:00 time right AI is not going to replace teachers okay AI is not going to replace writers because we need teachers we need experts we need writers to supervise the AI okay and that's what education should be for now okay and so I want to give this quote about AI literacy as you know as I mentioned before I study media and information literacy AI literacy is a large component of that now
• 27:00 - 27:30 okay um and this is the modern language Association and the four C's or the college composition and communication conference they had a Joint Task Force on writing and AI so they're really looking at what is the state of writing now and they said critical AI literacy is now part of digital literacy and students and teachers should be made aware of bias and inaccuracy in model outputs and the particular vulnerability of students who may not yet have sufficient expertise to critically
• 27:30 - 28:00 evaluate language model outputs including seeing them as sentient so what that means is this is the new skill that we need to cultivate in education and as students these are the skills that you should be asking for right how do I develop critical AI literacy to be aware of bias to critically evaluate what is the outputs of the language models and also to be critical of seeing these models as sentient I think
• 28:00 - 28:30 it's easy to maybe assume that they're sort of human-like because of the way they can produce things like a human speech and text like a human but they are not humans okay so you know whenever you can turn to your neighbor talk to your friends your family and remember what it means to talk to humans because human intelligence is still surpassing is still more advanced complex than machine intelligence in many ways okay and I have this quote
• 28:30 - 29:00 from another major thinker in artificial intelligence Stuart Russell he says humans are intelligent to the extent that our actions can be expected to achieve our objectives so as you go through your education ask yourself what are my objectives what do I want to learn how do I want to be in the world to be successful you know if you're in education you know in school just to like jump through hoops and get a degree get the piece of paper then
• 29:00 - 29:30 um you know okay you're going to use chat gbt to write your essay but if you're here to actually learn to actually be a better functioning member of society then you know that objective will lead you to doing different things with your education to learning in a different way to really appreciating the new lessons that um that your education has to offer okay and as you are learning new things I think it's really important to focus on metacognition which is a sort of higher
• 29:30 - 30:00 level thinking it's self-reflection and thinking about your own thinking okay if AI if new language models can produce can summarize texts what it can't do is reflect on why it did that you know it cannot reflect on how it got to the answer that it did because you'll maybe hear of AI talked about as this black box which means it's this sort of like you know weird Black Box where a bunch
• 30:00 - 30:30 of magic happens and then suddenly an essay gets spit out but it can't understand why it produced that we can we can ask ourselves why did I write that why do I think that you know what did I experience what did I read that led me to want to say this okay and that is still what makes us human our ability to reflect on our own processes okay so as you're going through your education really think about that think about how
• 30:30 - 31:00 and why am I learning this and as you're learning it's important to focus on the process over the product okay the essay yes that's the thing you submit that's the thing you need to give your teacher but the process that you went through what you learned in writing it that is something that you get to keep and reflect on so that you can be a stronger you know more informed more metacognitive student and as a student as a human I encourage you to lean into
• 31:00 - 31:30 wonder and curiosity because one of the industries well okay if you want to think about it from a work perspective one of the industries that's developing is called prompt engineering which is where you basically have to you learn how to ask these chat Bots better questions right it's learning how to give it the right questions so that you get a more effective output but that is based on a very human capability of asking questions of trying to figure out what
• 31:30 - 32:00 is it that I don't know that I want to know and how do I phrase it in the right way so that I get you know kind of what I need from this tool and and again that's part of our human intelligence wondering and curiosity so really you know Embrace that remember that that is what makes us stronger and you know as you get better at asking questions then maybe you'll become a very successful prompt engineer and you can credit me for that actually um see okay and I think that's it so thank
• 32:00 - 32:30 you very much for listening to my presentation I hope that was helpful [Music]