Back
Jeffrey Maggioncalda
Former Chief Executive Officer, President & Director, Coursera

The Future of Learning and Work: How GenAI is Creating More Equal Opportunity

🎥 Sep 18, 2024 📺 Coursera ⏱ 67m 👁 3613 views
GenAI will unleash a new wave of reskilling imperatives worldwide, reshaping nearly every job role and company. Hear from CEO ...
Watch on YouTube

About Jeffrey Maggioncalda

Jeff Maggioncalda, CEO of Coursera, has been focused on the integration and impact of generative AI on education and the workforce. He has stated that generative AI is the number one in-demand skill on the platform, with enrollments in related courses increasing significantly. Maggioncalda has described AI as a "thinking tool" that will change how people make decisions and perform cognitive tasks. He has argued that companies are shifting from developing an AI strategy to implementing it, with a focus on teaching employees in specific job roles how to use the technology to improve productivity. He has also noted that human skills like assertiveness and communication are among the fastest-growing skills sought by learners. Coursera has introduced several AI-driven features under Maggioncalda's leadership. These include Coursera Coach, a personalized tutor and teaching assistant, and a suite of tools to prevent cheating, such as AI-based proctoring, browser lockdown, and a feature where an AI interviews students about their written submissions. Maggioncalda has also launched his own course, "Navigating Generative AI: A CEO Playbook," and has stated that "AI will not replace professors but professors using AI will replace professors who are not using AI." He has discussed the company's partnerships with universities and industry leaders to create job-specific training programs and micro-credentials, and has emphasized the importance of lifelong learning as the rate of technological change accelerates.

Source: AI-verified profile updated from Jeffrey Maggioncalda's recent appearances. Browse all interviews →

Transcript (50 segments)
J
Jeffrey Maggioncalda0:00
Please welcome to the stage Coursera's Chief Executive Officer, Jeff Maggioncalda. Thank you, thank you, thank you. Good morning. How many people had a good time last night in Las Vegas? I mean, pretty amazing town. We're going to start off today with a keynote talking about the future of learning and work. The future of work is going to have a big impact on shaping the future of learning. As technology advances and globalization occurs, businesses are going to be changing the way that they work, functions will be changing the way that they work, jobs will be changing, and the knowledge, skills, and tools that people use to do those jobs will be changing as well. So people will have to learn more quickly as the world changes more quickly. I like to say that the rate of learning must match the rate of change, and with generative AI, the rate of change is accelerating. As Alondra and as Marne mentioned, the disruption that could come with Gen AI could threaten to widen inequality, could potentially have a lot of biases and harms to different populations. Some would say it could even potentially threaten civilization. But just as Gen AI is one of the biggest threats that we're facing and is causing the fastest change, Gen AI is also one of the greatest enablers to learning that we've ever seen. So we're going to talk a little bit about the future of learning and work, but rather than imagining some distant future, we're going to talk a lot more about what's happening right now, what's happening with our product, what's happening with our customers, what's happening with our partners. This is going to be a real exciting show and tell. I brought some friends along, and they're going to be showing you a lot of what's possible in this new world of generative AI. I'd like to start by just thanking our sponsors AWS, IBM, Microsoft, and Google Cloud for supporting this event and for bringing us together.
I like to ground this topic in the United Nations Sustainable Development Goals. This is basically the recipe that nations put together in 2015 as a formula for creating human prosperity and for protecting our planet. I know I speak for many of you in thinking that one of these stands out above the rest: SDG 4, education. Because education is what moves humanity forward. It's the thing that helps us understand and manage our climate, it's the thing that helps us understand and manage our health, it's the thing that can reduce inequality and build strong, just institutions. It's a reason why these two people, Andrew Ng and Daphne Koller, founded Coursera in 2012: the belief that knowledge and education moves humanity forward, and we should do our best to make sure it's accessible to everybody in the world. What started as a collaboration of two professors turned into a collaboration of a handful of elite universities, and this collaboration expanded to universities around the world, and this collaboration expanded further to include industry leaders around the world, all brought together with the common purpose of creating universal access to world-class learning. We made tremendous progress, and I think we're just beginning. Andrew said back in 2016 that AI is the new electricity. As Dr. Nelson said, it's a general purpose technology like electricity. We don't know where this is really going, we don't know the impact and implications of this technology because we don't know fully what the applications of this technology are. But the best way to find out, to understand the opportunities, to understand the threats, and to mitigate those threats, is to start using it in a responsible and collective way now.
I've had the privilege as the CEO of Coursera to go all around the world in the last couple years. I've talked to leaders in businesses, I've talked to leaders in government, I've talked to leaders on campuses. Everybody is absolutely enthralled, to some degree fearful, in many cases optimistic, by the possibilities of generative AI. Countries and leaders of countries are very interested in the potential economic opportunities that could be brought to their countries because of the possibilities of generative AI. There's lots of expectations that GDPs could grow considerably more in certain regions than others. Business leaders are very interested in generative AI. According to McKinsey, productivity gains stand to be above $4 trillion. And if you think about where the productivity gains are going to really occur, they're in certain functions: in customer support, in software engineering, in product and R&D, in sales, in marketing. Just those functions alone stand to account for more than $3 trillion of productivity gains. You can imagine that almost every CFO and almost every CEO is thinking about how to train their people to use these new tools, to think differently and work differently, to be more effective in creating value for customers. Individuals are also very interested in generative AI, partly because there are so many changes, good and bad, to the economic opportunity set available to them. According to Indeed, job postings for Gen AI skills are up 20 times. When Microsoft and LinkedIn interviewed leaders, over 70% of them said they would rather hire a candidate who has Gen AI skills and very little experience than a candidate who does not understand Gen AI and has a lot of experience. This could really change the experience set in the global labor markets. And then when Amazon did its study, it shows that people who have Gen AI skills often command a wage premium of 30 to 40%. So yeah, individuals are pretty interested in generative AI.
Now what does that mean for us at Coursera and in the Coursera community? It means we have a lot of work to do, and we're working hard to take this technology and put it to good use. As I mentioned, in the early days, Coursera and our partners started creating courses. We now make courses available to more than 150 million learners around the world. Today we have more than 300 education partners, universities and industry partners, who author more than 7,000 courses and credentials and offer them on Coursera. They offer them not only to individuals directly, but also to institutions. We have Coursera for Business, Coursera for Government, and Coursera for Campus. We'll be sharing case studies about how these institutions are using the content and credentials on Coursera. Now the way I'm going to frame out this keynote is I'm going to walk through these three basic pieces: the what are learners doing, what are we doing with our educator partners especially to use this technology, and then we'll go through some case studies about how institutions are putting this to use to serve learners in their institutions. All right, so let's start with learners.
Learners have really been benefiting from the work that we're all doing. I think it's a real testament to this community that by the end of this year, at current rates, we expect more than 160 million people around the world to have access to the content that's being created on Coursera. Course enrollments are following a similar upward trend. Now what's driving a lot of this is individual interest in generative AI. From a start of zero basically in Q3 of 2022, we now have almost a million Gen AI enrollments in just Q3 of 2024 alone, more than three million enrollments in Gen AI content on Coursera. Now we're getting about six enrollments per minute in generative AI titles on Coursera. So there is a real demand out there to learn about this. People know that the world is changing, and the rate of learning and the ability to learn these skills will be very important. Our partners have been very busy. They have created more than 500 generative AI titles. These are the most popular ones on Coursera so far. University partners from Vanderbilt to University of Glasgow, industry partners from Google to Microsoft, are creating a wide range of content and credentials to teach people what this means and how to use it responsibly. Where are these learners? Of course they're spread around the world, but notice: 26 million registered learners in India, more than all of Europe combined and growing at a faster rate. 25 million registered learners in Latin America, equaling the number of learners in Europe and growing at a faster rate than both Europe and North America. And the region with the fastest growth rate in the whole world is Africa. This is the job of human capital development. Where is the future of human capital going to be? Where the young people are. That's a big part of our job: to make sure that everybody has access to this kind of learning at a time when learning has never been so important. I'm excited to say that more than half of the Gen AI enrollments are from learners in emerging markets: India, Pakistan, Brazil, the three countries with the highest number of enrollments in Gen AI content.
Let's hear for a moment a story of one of the learners and how his life has been changed by some of the work that we've been doing, in particular courses that have been developed at Vanderbilt University.
N
Narrator10:47
Describe scene: opening scene, a low-lit interior space with filming equipment and three individuals, possibly preparing for a video production. Large windows in the background overlook a street, contributing to the room's backlight.
L
Learner11:06
The Coursera course that Jules taught was really transformative. Not just learning how to effectively AI prompt and how to get the information I want, but applying it to all kinds of different aspects of my life, like the AI art I do and creative story writing I've been doing as well, and getting a new job. I mean, I basically went from working a low-level state job to working for an AI company.
J
Jeffrey Maggioncalda11:50
It's easy to lose sight in these huge numbers of the individual lives that are changed because of the work that we're doing together. AI is a disruptor, but AI is also one of the greatest enablers in education that we've ever seen. The way that we think about it at Coursera is pretty simple. AI will be changing jobs. It will require a level of upskilling and reskilling at a rate that we've never seen before. As a company and as a community, we have to respond to that. So when we think about upskilling, people learning new skills for their current job, we and our partners have launched Generative AI Academy. This is teaching people about generative AI. We'll talk about this in the keynote coming up. We also have a lot of people whose jobs will be so significantly impacted by this technology that they're going to have to get new jobs and start new careers. We have another set of content and credentials from our partners called Career Academy. It helps train people, whether you're a student just starting a career or whether you're in a career and want to switch careers. It starts you from scratch to learn the skills of a new career. Not only will Gen AI change jobs and the knowledge you need, it will also change the way that we learn and the way that we teach. I'll give you an update on our machine translation efforts, which we talked about at last conference. We'll talk about Coach, which is an AI-based tutor who's doing incredible work right now. We have plans and we'll show you where we're heading in terms of providing more learner support on a more personalized basis. We'll also show you not only how Coach can be a tutor, but how Coach can also be a teacher. In the hands of a great instructor, Coach becomes a tool for personalized interactive instruction. We'll also talk about Course Builder, which is the ability to tailor courses to be specific to the needs of a certain learner audience. I want to point out that on Coursera, our partners are not just creating content. These are world-renowned institutions that also issue credentials. Credentials are signals that someone has mastered a certain set of knowledge or skills, and you want to make sure that that mastery is authentic. This is why academic integrity is very important. Credentials mean nothing if the learning isn't authentic. We're going to show you some things today about how Gen AI can make sure that that learning is more authentic by detecting and deterring student misconduct. Cheating with ChatGPT, Chai is going to talk about this, and I think you're going to find it pretty striking.
First on AI translations, quick update. I'm excited to say that since last year when we talked at conference, we've doubled the number of courses that have been translated and we've tripled the number of languages that these courses have been translated into. We talk about universal access to world-class learning. Most of the courses on Coursera are in English. Most of the people in the world do not speak English. The language barrier has been one of the biggest limitations holding us back, and that barrier is falling rapidly. What used to cost $133,000 to translate one course into one language now costs $20 to $25. We have millions of learners taking these translated courses and getting access to learning that they just never had before. Now I'd like to call up to stage Alice, one of our product managers. Alice actually was the product manager that worked on the language translations. Hello Alice.
A
Alice15:21
Hello, nice to see you.
J
Jeffrey Maggioncalda15:29
I'm just going to compliment you a little bit, Alice. Alice actually worked on the language translation feature, she continues to, but has shifted her focus and is doing a lot more work on Coach. Yes, Alice, what are you going to show us with respect to Coach?
A
Alice15:44
If you remember last year, we introduced Coursera Coach in order to make online learning more engaging by using the power of generative AI. The goal was to create a more personalized and interactive experience that is grounded in the trusted and expert content of our partners, but also developed using Coursera's pedagogical best practices. Now today, we'll show you how Coach has transformed how learners learn on Coursera, but also we've given Coach a bigger mission: to turn our learners into achievers. So let's take a look. What you see here is a typical video lecture on Coursera, but what is new is that we've made Coach available at any point in time during their learning process to help learners get unstuck or to help them if they run into some challenging concepts. Since making Coach available in the hands of the learners, we've observed how they find ways to make Coach even more useful to them in the context of what they're learning. Here's an example: learners might ask, 'How does this concept I'm learning apply to my job?' Coach will come up with examples that are particularly relevant to their role and in a way that will make it a lot easier to remember all the concepts that they're learning. Another use case that we've seen is that as learners make progress through a course, they tend to drop off right before a quiz. Coach is actually perfect to help them practice what they've learned before they move on to the quiz. In this type of Q&A format, Coach will help the learners recall the key concepts, but also will give them pointers and feedback so they know exactly what they grasped well, but also what they may need to revise further. At the end of the session, Coach will recap the key concepts so that again they can reinforce the material. This type of interactive learning is really what is giving learners the confidence to take the next step in their learning. So far, Coach has already helped over a million learners make progress and stay on track. Those who do use Coach are more likely to pass quizzes on their first attempt. When we talk to them directly, they continue to tell us how Coach has been a game changer because it has made it a lot more efficient, but also has given them the opportunity to have a one-on-one interaction with a peer or a teacher. That is the type of more personalized interaction that was unimaginable in the past, but now we can make it possible.
J
Jeffrey Maggioncalda18:11
All right, Alice, you teased us a little bit by saying that you're expanding the mission of Coach to do more to help them turn into achievers. What do you mean by that?
A
Alice18:23
We've taken a massive step in the evolution of Coach, and that is in guiding them throughout their career discovery and progression.
J
Jeffrey Maggioncalda18:28
Okay, yeah, so let's take a look.
A
Alice18:34
This is Coursera's homepage, and what you'll see is we're highlighting Coursera as a destination where people can come and grow their careers. Learners will be able to explore over 40 in-demand roles. For people who may not know which careers to start with, Coach will be available to help them narrow that down. Now what we've learned is that most people who come to Coursera actually have an idea of one or two different roles that they may want to pursue. So let's imagine I would like to become a data analyst because I love problem solving and I'm pretty good with numbers. I can drill into what does it entail to become a data analyst, and I can see that Coursera is offering a variety of programs to help me get there that are tailored to different experience levels. But it might still be a little bit difficult for me to understand which particular credential is right for me. So I can ask Coach to help me find the best recommendations. Through this series of questions, Coach will get to know me better and my goal. Here I can share that I want to switch careers to become a data analyst, and Coach may ask, 'Well, what am I doing today?' I'm a researcher. Now based on this information, Coach will start to understand what type of transferable skills I bring to the table that I can build upon in order to switch careers, and it will refine the recommendations for me. Now Coach has given me a short list of the programs. Essentially these programs are made up of a series of courses that will allow me to progress and build the skills in order to become a data analyst. If I want to understand why this particular program was recommended for me, Coach will give me a rationale that considers what he has learned about me in our previous conversations. Let's say I actually want something a little bit more advanced because I have programming experience. I can share that with Coach, and Coach will refine the recommendation to find something that is actually more tailored to not only my goals but also my experience level. With this new experience with Coach, we're really excited to help our learners not only find the right career for them, but also have the right path to be able to get there.
J
Jeffrey Maggioncalda20:41
All right, so clearly Coach is out there helping lots of people. You've just shown us something that is not yet available. When is this new career guidance going to be available to learners?
A
Alice20:51
We are currently testing, and it should be available in the hands of the learners by the end of the year.
J
Jeffrey Maggioncalda21:01
Okay, great. Thank you, Alice. Great demo. So that's a little bit about how we're using Gen AI to provide learner support as a tutor and for career guidance. Let's shift now to educators, because not only will we be helping learners learn in a different way, we'll be helping teachers teach in a different way as well. As I mentioned, there are more than 300 institutions who are partners with Coursera, putting content and credentials on the platform. On the university side, our scope and range of university partners continues to grow. Since last conference, we have announced eight new university partners, leading universities from around the world. Today I'm excited to announce four additional university partners joining the Coursera community: Saïd Business School at the University of Oxford, IMD Business School, Real Madrid Graduate School, and the Dubai College of Tourism. On the industry partner side, we continue to grow our industry partners very rapidly. Since last conference, we have added 20 new industry partners, you could see in different industries and across different regions. I'm excited to announce today five new industry partners joining the Coursera community: Airbus, Amazon, Johns Hopkins Medicine, Liberty Mutual Insurance, and Xbox. Clearly a wide range of industries, regions, and job roles will be assisted by the content and credentials from these new partners. Together, we want to make it so that any learner in the world, no matter who they are, no matter where they are, no matter what language they speak, can go to a place and get access to the highest quality, most trusted, authoritative expertise from both higher education and from industry.
Now let's talk a little bit about Generative AI Academy. We call it the Gen Academy for short. We launched it in November of 2023. We started with AI for Everyone, a series of courses that are for the general audience, things on what it is, how it works, and how to use it responsibly. Then in the early part of this year, we launched Generative AI for Executives. This is for business leaders who are really trying to navigate this landscape that's changing so quickly. Jules White from Vanderbilt also created a course called Gen AI for University Leaders to help academic institutions navigate the changes associated with Gen AI. More recently, just in the last couple months, we launched Gen AI for Teams. This is a set of content and credentials that really gets to specific job functions and job roles. The way that a software engineer uses generative AI to do her job is very different than other jobs. Software programming is a language, large language models are quite good at programming. The efficiency, the quality, and automating the tedium of many tasks associated with software engineering are changing rapidly, and this content really helps software engineers and product managers utilize these new capabilities to do their jobs well. It's different from data analysis. There are a lot of new techniques and tools for using Gen AI as a data analyst, and there's a whole set of courses and content to help data analysts be more productive, efficient, and work with higher quality. We also have Gen AI for Marketing Teams. Marketers deal with media, with language, with images, with sounds, with videos. This is precisely the kind of stuff that generative AI excels at, so marketers' jobs are changing dramatically with the rise of these tools. But the way that a marketer uses these technologies and the way a software engineer uses these technologies is quite different. We're going to be marching through every job function in companies, helping folks understand the technology, how to use it, how to use it responsibly in order to do a better job, and ideally create job roles that automate the tedium and make jobs more creative, more engaging, more valuable, and ultimately more satisfying.
Shifting to universities, we just completed a survey of a thousand university leaders around the world. When we think about this Gen AI in business, we're seeing a lot of business and industry content become very interesting to leaders in universities, in particular professional certificates and industry micro-credentials. In the survey, 94% of university leaders said that they believe that industry micro-credentials could strengthen the skills and the likelihood that their graduates get jobs. Over 50% of these leaders said that they already offer industry micro-credentials to their students to supplement the curricular offerings of the university. Among these, more than half offer industry micro-credentials for credit towards a college degree. So for all the people who say that universities are too slow, that degrees will become irrelevant, they're failing to realize the power of adaptability and collaboration. Universities are integrating industry content into their curricular offerings to the benefit of their students, their faculty, and the employers who hire their graduates. This is happening at a faster and faster rate. Now what do these industry micro-credentials look like? It's these professional certificates that have been created by some of the best industry leaders in the world. It started in January of 2018 when Google launched the IT Support Certificate on Coursera, and you can see that catalog has grown by job role, by function, and by partner who's authoring it. We've been working with credit recommendation agencies, the American Council on Education in the United States and the European Credit Transfer System which is the framework in Europe, so that these certificates can be reviewed for rigor and college standards and be more easily integrated into the college curriculum. We also recently received designation as an ACE Authorized Instructional Platform, which is broadening from just reviewing content to actually reviewing and setting standards for technology platforms, especially as we enter the age of AI. I'm also excited to announce today that we've been working under India's new education policy to make sure that these professional certificates meet the standards of the National Skills Qualification Framework in India. It's not just the US and Europe. In India, they are basically creating these quality standards to facilitate the integration of industry micro-credentials into college and university programs. This is happening around the world. There's a collaboration that we just have not seen before between industry and higher education, and I think it is really something that is growing and growing more quickly. To date, we've already had more than 12 million people enroll in these professional certificates. I can assure you they come from the best experts, and learners are eating them up, and no learners more than students in college wanting to understand what it will take to be ready for a workforce that looks vastly different than it did even three, four, five years ago. Since last conference, we launched 29 new professional certificates across many different industries and from many different providers. I'm excited to announce today seven new entry-level professional certificates from ADP to Microsoft to Epic Games for video game designing. These are the jobs of the future. I'm also excited to say that not only are our partners creating new titles in this new world of Gen AI, they're refreshing their existing popular titles and integrating generative AI concepts, knowledge, and techniques into those titles so that learners taking these will be ready for what's coming next, not necessarily just what was the skill set and the knowledge that was required in years past. I'm also excited to announce today that Google has upgraded all six of their major entry-level career certificates with generative AI powered by Google Gemini. So now when people come to Coursera and take these professional certificates, they will be trained to do these jobs in a world where these jobs are heavily assisted by generative AI. Things are moving fast. Our partners are moving fast as well.
Let's hear quickly from one of our Coursera for Campus customers, Louisiana State University, to share how they are looking at these industry micro-credentials as part of their college degree programs.
R
Radhika Krishnadas30:15
My name is Radhika Krishnadas. I'm the Assistant Vice President of Program and Learning Design with LSU Online and Continuing Education. In today's learn and earn ecosystem, it is very critical for a higher ed institution to offer job-relevant, market-ready certificates that would help students to become more employable and future-equip them as professionals. Coursera has a really comprehensive portfolio of professional certificates in very in-demand areas offered by the leading industry giants. We saw that most of them were ACE credit-recommended, which meant that several subject matter experts from around the world have vetted those credentials for quality, for rigor, and for meeting college standards. So we were able to accept those and also able to give credit for prior learning, so that whoever took a professional certificate could apply those credits from those certificates towards a future degree.
J
Jeffrey Maggioncalda31:19
We're seeing a lot of innovation in college degrees. I also just want to point out that many of our university partners offer degree programs on Coursera, bachelor's and master's degree programs all around the world. Since last year, there have been 11 new degree programs that our partners have put onto Coursera. There's a new program we'd like to announce today with the University of London. It's a foundation program for their Bachelor of Computer Science for people who maybe didn't have a background in computer science who want to switch into this but don't have the prerequisites. This is a one-year online program that provides a pathway into the field of software engineering. Now I'd like to talk a little bit about using Gen AI to help educators. I think you'll find that the world looks vastly different than it did even six to 12 months ago. We've talked about Coach as a tutor to help learners. Now we're going to hear about Coach as an instructional partner to extend the instructor's ability to offer personalized, interactive instructions to individuals. To really present this, I'd like to invite Nat Cooper to the stage. She's one of our designers who's actually been building these Coach dialogue items. Nat, welcome to the stage.
N
Nat Cooper32:48
Thank you. All right, so for me, this is sort of the highlight of the keynote. What are we going to see today with respect to Coach in the hands of instructors?
J
Jeffrey Maggioncalda32:54
Sure, yeah. Thanks, Jeff. Online education itself can present unique challenges, from delivering personalized instruction at scale to keeping learners engaged and motivated, to providing them with timely and relevant feedback that supports their progress. So that's why we're excited about Coach's potential to improve learning outcomes through personalized interactions. Now instead of just talking about it, let me show you how educators can use Coach in their courses to be a transformative tool to elevate online learning. All right, let's take a look.
N
Nat Cooper33:40
All right, let's start the demo. As an instructor, you may want to bring the interactive and dynamic elements of your physical classroom into the online learning environment. As you build your course on Coursera, you now discover that you can add hands-on learning activities facilitated by Coach. These activities, which include roleplay and dialogues, enable interactions that were previously limited to face-to-face settings. Let's explore one of these activities together. We're going to take a look at the dialogue activity today. Note that the dialogue comes preconfigured. It comes with a mentor persona, which is one of many personas you can choose from, all designed for different educational contexts. It also recommends different interaction modes such as text-based and even voice exchanges. Remember, you're in control here. You can define the purpose of the activity and rubric details, which will provide instructions for Coach on how to conduct the dialogue itself. The activity incorporates all course material that you've included up until this point for context, along with Coursera's pedagogical best practices and learning design principles. This not only enhances the relevance and the quality of the dialogue, but it also improves the accuracy of the AI questions and responses. Now before adding this to your course, you'll want to view it as a learner to see exactly what they'll experience. Following your expert guidance, Coach will start to ask the learner probing questions, encouraging them to really stop and think and develop their own ideas. When prompting learners to reflect on their responses, Coach will draw on the materials that you, the educator, provided in order to help deepen the understanding of the topic. Now rather than just texting with Coach, what if I could actually talk with Coach? Ah, yes, this feels much better. Where were we? We were talking about effective communication with clients. What methods of communication can you use to understand the client's needs and limitations? The shift from a text-based exchange to a full-on conversation really enhances that experience with Coach. Now it feels like learners can interact directly with their instructor. Our world-class faculty haven't just been using Coach in their courses, they've been instrumental in helping to shape them with their expertise and guidance in our development process. You know that these activities are designed by educators who hold the highest standards in teaching and learning.
J
Jeffrey Maggioncalda36:23
All right, Nat, this is pretty amazing. I would like to first point out Barb Oakley somewhere here in the audience. Barb just launched a course called Accelerating Learning with Gen AI, and it has these Coach items in there. Where are we in the actual deployment of this, both the text space but you also mentioned audio?
N
Nat Cooper36:45
The audio mode is underway, and I'm excited to announce that the text-based mode is live today, used by real learners in real courses today.
J
Jeffrey Maggioncalda36:57
Nat, thank you. Pleasure. So I actually put one of these dialogue items in my course too. There's a course called Using Gen AI as a Thought Partner, a piece of which is your role as a critical thinker. After the video, Coach comes up and challenges your critical thinking in the course itself. But someone who's been using this and shaping it much more than I is a special guest that I'm really excited to invite to the stage, my friend Vic Strecher from the University of Michigan. Hey, Vic.
V
Vic Strecher37:30
Thank you so much.
J
Jeffrey Maggioncalda37:39
Vic, this has been such a thrilling adventure so far. I mean, what you've done with neuroscience and what you've done on Coursera with your courses has been extraordinary. Vic has a couple courses, a quarter million people or so have taken them on finding purpose and meaning in your life. You also offer an app so that people can develop habits to reinforce their sense of purpose in their lives. Tell us a little bit about what you've been doing with our product team and as an instructor using Coach items.
V
Vic Strecher38:02
Well, first of all, thank you. Thank you, Jeff. Thank you, Coursera, for helping me reach more people than I have in my 40 years of teaching. That's no joke. It's really fulfilling as a teacher to be able to do this. One of my purposes in life now at this stage is to help my students, my learners, develop greater purpose and intentionality in their lives. So this is really helping them do that. It's been a lot of fun working with your team as we've built in this amazing Coach, this Socratic dialogue that we can interact with learners. Could you show us a little bit how you're using this in your courses?
J
Jeffrey Maggioncalda38:39
Oh, okay, all right, sure. First of all, what's very cool about this is that Coach has already been exploring my course. It goes through my videos, it goes through all the teaching that I've had up to this point, as Nat was saying. So here, this is what I put in. I just say, here's the purpose of this particular activity. This is a dialogue. By the way, I have 11 dialogues that I've put into my Finding Purpose course. The first one may have taken a half an hour, then I did three of them during the halftime of the University of Michigan football game. It's not so hard to continue putting these in and start authoring these dialogues with this Coach. In this particular example, I put in, I just wanted to kind of a warm-up exercise: knowing the difference between having a purpose in your life and actually being purposeful, aligning your behaviors, your emotions with that purpose. I put that in. Right away, Coach develops milestones. The milestones are very smart because it already understands my course, right? So it's now generating smart milestones for me, and it asked me, do I want to start? Yeah, so ask permission. That's really an important part. Say sure, great, let's go ahead. Think about someone who has a clear purpose in life. So they start asking me a question related to purpose versus being purposeful, and I answer the question. It gives me some very positive feedback right away: 'Great observation.' And then it elaborates more on my answer, which I loved. So I start really becoming engaged. Then it asks, 'Can you tell me what comes to mind when you hear the phrase life purpose?' Well, I teach a course to many, many hundreds of students around finding purpose in their lives, and unfortunately, many of those people, and many of our students, I'm sure a lot of you can relate, have depression, have high anxiety, have big issues right now. So I decided to get real, and I said, 'I don't have a purpose. I tried to channel a student of mine. I feel lost and I feel depressed.' How does the Coach respond to that? 'I understand you're feeling this way.' It's empathetic, but at the same time, just as I would be doing in my own class, I try to steer that student back to the purpose of the course, and that's what's so important. Then it essentially asks whether it can start giving me some examples and some ways of thinking about purpose a little more. I go, 'Yeah, fine, sure.' And it says, 'Okay, well, let's imagine a professor.' By the way, as I've been testing this, it might say imagine a novelist or imagine a plumber. It doesn't matter. It comes up and it always surprises me, and that's something that's very delightful. This is not canned by any means. It comes up with new scenarios. So it asked about a professor: 'How do you think this professor's passion for inspiring students could translate into a sense of purpose for them?' I say, look, I decide to get real here because I am a professor, so I thought, wow, this is a great way to really become authentic and test this program. So I say, 'In my career as a professor, I want to teach every student as if they're my own child.' It gives me some very strong feedback, very positive feedback: 'You've touched on a crucial aspect of emotional alignment.' It goes on, and then it asks me a really tough question: 'Can you elaborate on a specific time when focusing on your purpose helped you navigate a difficult time, a difficult situation with more resilience and emotional stability?' I decided to get very real, and I said, 'When my 19-year-old daughter died unexpectedly of a heart attack, which she did, I needed to repurpose my own life. My new purpose focuses on helping my students, who each year are the same age as my daughter, and this saved my life.' And it really did. Coach gives me strong feedback, very positive: 'That's incredibly powerful and inspiring. Your story beautifully illustrates how aligning with your purpose can provide strength and direction.' Then another thing that I wanted to incorporate into this particular dialogue, of the 11 dialogues that are in there now, I wanted to explore different strengths that you might need to be purposeful. So it starts asking me that. What do I need to have to be purposeful in my own career as a professor? I found that one of my most valuable strengths is being inspiring, and I go on, and the Coach gives me feedback. Now you're finished, you're 100% complete. By the way, it says, 'You've demonstrated an advanced level of understanding,' which it sure better have said that, right? And it gives me the strengths, and it's really tailored to me, very deeply tailored to the things that I pointed out in my responses. If I had areas that I needed to grow in, it would tell me so. This has been an incredibly powerful tool for me. I will say this just to summarize: I think in my career, I'm so used to large classes with a screen or a whiteboard or a chalkboard or whatever, and I'm trying to teach in a one-way communication, one-way approach to all of these students, hoping that this one-size-fits-all reaches people. Obviously, now reaching 250,000 people, you know, that's the same thing, hoping that this would reach the right people at the right time. But now with this coaching dialogue, I feel like there's a third dimension to this. I feel like I'm reaching out to people almost as if in class, and they stay at the end of class and say, 'Can I talk to you? Can we have a coffee?' Or during office hours, and I can really have a dialogue with an individual student. Now I can reach 250,000 individual students with this type of dialogue, and I hope to be reaching hundreds of thousands and hopefully even millions of students in the future. Thank you.
You never cease to amaze me. This is the privilege of being at Coursera, working with people with your level of expertise, your heart, and your willingness to open up and connect with other people, and to serve them in such a profound way. Thank you, Vic. I really appreciate it.
V
Vic Strecher45:09
Thank you.
J
Jeffrey Maggioncalda45:22
So you might be sitting here saying, 'Well, where's Coach really going to go? How much of that was it following Vic's course and Vic's instructions and the context that Vic pulled in, and how much of that was the AI?' We don't know the exact boundaries right now. We're going to be monitoring it. We're going to be working with Vic and other institutions and other experts, and we're going to navigate it as we go. I will say though that we're generally erring on the side of controlling and constraining the kinds of things that could be done. Vic, the number of times that you said this, and Nat said the same thing, and we did this with the translations, and we're doing this with Coach's tutor, grounding the AI in the human expertise and oversight is critical. This is a thinking tool. It's an extension of the instructor, not a replacement of the instructor. That is really where we're going. One of our key responsibility principles at Coursera is accountability. We must be accountable for the way we use this and the impact it has. In a world where you're not sure exactly what that impact is, does it mean don't do it at all? In our view, no. It means do it responsibly, monitor the impact, and mitigate the potential risks that it could create harm to some population. But unless we start using it and start using it as responsibly as we can, we will not learn what the frontiers and boundaries might be. Now you've seen some pretty amazing things. I will say that and thank and acknowledge one of our most important partners at Coursera, Google. All of these Coach items at this point are powered by Google's Gemini Pro 1.5 model. As these models improve, as Nat said and as Alice said, it won't just be text, it'll be text and audio. It won't just be text and audio, it will be images. It won't just be text, audio, and images, it will be videos. The ability, once you have the fundamental concepts and grounding of the course, the instructors, the knowledge, and the general rubrics, the modality will be relatively straightforward. As these underlying models get better, more capable, your ability to instruct students, the ability for your faculty to instruct students, I think will be unprecedented. So that's a little bit about Coach as teacher in the hands of a teacher. Now I'd like to talk a little bit about Course Builder, which is generative AI to actually build courses that are tailored to the needs of a special population or a particular population. I'd like to invite Austin up to the stage to talk about Course Builder. Hello, Austin.
A
Austin47:50
Hey, Jeff. Thank you. Great to see you.
J
Jeffrey Maggioncalda47:58
So Austin writes software code. He's not a designer, he's not a product manager. He actually builds this stuff. That's true. Austin, what are you going to show us about Course Builder?
A
Austin48:07
Well, Jeff, as you know, Course Builder is an AI-powered authoring tool that helps you quickly and easily create custom courses leveraging the global expertise on the Coursera platform. Today I'm excited to share with you how Course Builder has evolved since we launched it during last year's conference, and give you a sneak peek at some of the new features we're building to make creating custom courses even easier. All right, let's take a look. All right, let's dive into the demo. To get started with Course Builder, we first have to select between one of two options. The first: building a brand new course, let us start from the ground up building a course custom tailored to our organization's unique needs. Alternatively, a new option we just launched last week is the ability to start with one or more existing courses, keeping what's relevant from them and removing what's not. Today we're going to focus on building a brand new course. To help the AI underpinning Course Builder align itself to our goals and structure our course, we can either use our guide to give it a little bit of information, or if we already have them, we can just upload some files. These files can be things like training recordings, instructional materials, or even a course syllabus if we have one. Once we upload these, Course Builder will analyze these files and extract for us some structured data, things like the course topic, our learning objectives, and who might be an appropriate audience for our course. It'll also give us a full outline of our course, including all the modules, their names, descriptions, and their learning objectives. If we want to iterate further, Coach can help. Course Builder Coach acts as an AI-powered authoring assistant. He can do things like refine content or suggest potential modifications to our course. For example, if I want to make that first module just a little bit more interactive, I can just click 'expand this module' and tell it to include some hands-on practice for us. Coach will then analyze that, regenerate the topics of our module, update its learning objectives, and give us a couple of options. We can look through these and figure out which version of the module we like the best. This highlights one of the core principles of Course Builder: the author is always in charge. Not just during your interactions with Coach, but in all of Course Builder, anything you see on my screen can be edited or regenerated at any time. Once we're happy with this outline, we can move to the next step where Course Builder will create a draft of our course. This draft will include a variety of proposed content, including some that it's found from a participating partner, as well as some content that Course Builder itself has generated for us to help improve the cohesion of the course and smooth out transitions. You'll also notice some of our uploads from earlier have been turned into videos and readings. Of course, Course Builder will include interactive activities like the dialogues you just heard about from Nat and Vic. When we're happy with the content of our draft, we can move forward to measuring the learning in our course using custom-generated assessments like final exams and module quizzes, both with built-in academic integrity features. When I click 'add assessments,' Course Builder will generate these assessments on the fly and embed them in pedagogically relevant points throughout our course. If I want to see the settings, I can click in and see the number of questions, their question type, and mastery level. If I want to dive in even deeper, I can click preview and see the exact questions that Course Builder has generated for us. Now that we're done with assessments, we're ready to launch our course. Just like that, we're live to our learners. We started with just a course concept and a handful of files, and now we have a live course with content from Coursera's vast catalog, interactive activities, and assessments with integrity. Since launching in beta 6 months ago, we have seen a sea change in organizations authoring custom content on Coursera's platform. In fact, we've seen over 715 new custom courses launched live to learners. Those have generated over 137,000 enrollments globally.
J
Jeffrey Maggioncalda52:06
Austin, pretty amazing stuff. So clearly Course Builder is out there, courses are being created, learners are taking these courses. But you also showed some stuff, which one of the things I love is Coach as instructional designer helping the instructor. When's that going to be live?
A
Austin52:25
You can look for a Coach for authors in the new year. We're not just focused on features right now. One of the key focuses of our team is expanding access to Course Builder. To that end, just last week we enabled it for organizational admins to self-serve turn on Course Builder for their users, and we're hard at work right now expanding access to Course Builder for our university partners.
J
Jeffrey Maggioncalda52:43
Awesome. Thank you, Austin. Thank you. Take care. So Austin said something pretty important. It's a theme throughout all of these demos: the human is in charge. This is a thinking tool. The accountability must lie with the person who's using it and the people who are creating it. Ultimately, though, this can help us all do our jobs far more effectively. Now Austin also showed you those assessments that get generated. He showed the exact questions. The author, by the way, can delete, edit, add, or change virtually anything in the course, because when they launch it, it's not AI that's responsible, it's the instructor who's responsible, it's the institution who's responsible, it's Coursera who's responsible. It's a huge part of what this is about: being accountable and responsible for the way that we use this technology. Doing assessments is great, but if students cheat, it doesn't mean much. We have seen a huge rise in students using ChatGPT to cheat. There are a lot of questions of what to do. I think most institutions have said, well, they're going to have to be using these tools in the workplace, so we can't say don't use them. But how do we detect and deter student misconduct so that the credentials that are issued matter, that they mean something, that the student and the employer who hires them says this really does reflect that you've mastered certain knowledge and skills? Well, to talk about academic integrity, I'd love to welcome Chai to the stage. He's going to talk about some of the new features we've been building with Gen AI. Chai, come on up.
C
Chai54:09
Nice. How's it going?
J
Jeffrey Maggioncalda54:20
Good, good to see you, Jeff. Okay, so this is like a circus of generative AI capabilities, and I think you might be showing something that people aren't really expecting. What are you going to show us today when it comes to academic integrity?
C
Chai54:33
Yeah, absolutely. Well, academic integrity can be a very complex topic in the age of Gen AI, for sure. But our strategy here was developed in deep collaboration with our own teaching and learning experts, leading pedagogical experts in the field, as well as educators from our partner and customer institutions. So everybody in the room here, there's a lot of pedagogical firepower here, and we really have a number of tools to achieve these goals. So why don't we take a look? Sounds good. Okay, I'll get set up here. Cool, let's start the demo. I'll start by navigating to the new Academic Integrity tab here, where I can set up course profiles for my credit courses. Any features I enable in this profile will automatically be enabled to courses applied to the profile. Features like graded item locking, which blocks access to graded assessments until all prior content in the course is completed. That can include things like videos and readings as well. Also, Coach blocking, so learners can't ask Coach for answers during timed graded assessments. For example, in this course, I'll enable both of them, and then I'll move on to assessments. I can also create assignment profiles. For example, if I have a closed-book final exam, something like a multiple-choice exam that has really high grade weight in my course, I can enable locking browser, which will block access to websites and applications on the learner's test-taking device, or something like Honorlock proctoring, which will go one step further by identifying resources within the learner's test-taking environment as well. Because this is a final exam, again quite high stakes, I'll enable Honorlock for maximum deterrence, but I'll also pair that with time and attempt limits to further reduce opportunities for misconduct. We've seen that increasingly, especially with what Knack and Vic showed as well, there's more and more desire for open-ended, project-based opportunities as well. So we've also built features for this use case. Features like plagiarism blocking, which will automatically block submissions that are too similar to other peers' work. I'll enable that. But in the age of Gen AI, educators are facing a brand new challenge right that goes beyond just plagiarism. Learners can generate long-form submissions in a matter of seconds. We've developed AI Via, one of Coursera's latest AI innovations, to help with this challenge. Our inspiration for this feature is deeply rooted in history, because educators since Socrates have been using live oral exams to probe out a learner's thought process in order to gauge mastery. We've developed AI Via with that in mind, where we want to scale this concept by providing personalized follow-up assessments to all learners. So I'll enable this, and let's dive deeper to see how it works. Now I'll put my learner hat on, where I'm being asked to submit a long-form business proposal. After I submit it, we'll see a dialogue start where AI Via will essentially review a learner's submission and then generate personalized follow-up assessments for them specifically through the dialogue. It can test up to five topics to gauge for authenticity. Now after the submission, AI Via takes the responses and evaluates them against a set of criteria. So it can include: did the learner show knowledge of and consistency with their original submission? Could they speak to the actual work that they submitted, as well as the depth and the relevance to the questions asked? So were they able to answer AI Via's questions well? At the end, AI Via generates an authenticity score. Now I'll take my learner hat off, put my educator hat back on. I'll go into the grading interface here, where I can see that AI Via automatically flags low scores, which indicates that a learner was unable to demonstrate a solid understanding of their own work. I can come into the AI details, the analysis details, to see the evaluation against the rubric and basically try to determine the next best course of action to support this learner. So we've seen a suite of tools here, which all work really towards one common goal: to empower educators to uphold academic integrity and ensure authentic learning at scale. Now taking a look at some of the impacts of these features, year to date we've seen over 6 million course completions go through with at least one academic integrity feature enabled. As a result, we've seen a 75% reduction in low-effort course completions, an 80% reduction in peer review plagiarism, as well as a 90% reduction in actions like copying, pasting, and navigating away from the exam window when a solution like proctoring has been enabled for a high-stakes exam. So back to you, Jeff.
J
Jeffrey Maggioncalda59:17
Great stuff. So there's a lot of stuff here, and much of it is live today, some of it's coming soon. What is live and what's coming and when?
C
Chai59:28
Yeah, I'm super excited about AI Via, which will be available later this year. But looking beyond to 2025, we want to continue pushing the boundaries of assessment innovation. You mentioned video before. We definitely want to add voice and video to AI Via to get closer and closer to the depth of a live oral exam. The closer we can get to that, the richer set of signals that we can provide educators to not only uphold academic integrity, but to improve learning outcomes, which of course is all of our main goals: to help the learner.
J
Jeffrey Maggioncalda59:54
Great. Chai, thanks a ton. Awesome. Thank you. So that's just a taste of what's going on. There are a lot of things that we didn't have time to show you today. Some of it is actually happening now, as you can see. Some that's coming in the next months. But we tried to really focus in on what we're working on that is happening or will be happening very, very soon. We've seen on the content side: Gen AI Academy and Career Academy content catalogs and credentials to help people adjust and learn the skills of a new workplace. We've also seen with translations and with Coach helping support learners, also with Coach helping extend the coach as a teacher, so having AI be an extension of the teacher to be a more personalized interactive instruction. The content is part of it, but the credential really matters, and knowing that the credential is valuable means that we need to make sure that especially in this area of online learning, which is growing faster and faster and faster, we can make sure that the learning is authentic. So that's a sample. I will mention that out in the hall there will be demo stations with product managers and designers, so you can put your hands on these things, check them out, see how they work, talk to the people that built them, and hopefully give us some ideas about where you'd like these things to head. Now the final piece to wrap up the keynote is institutions. I would just love to give you a couple of examples of how this technology, these content, and these credentials are being used in the real world. We work with companies all over the world, as well as universities, as well as governments, in many different industries. I thought I would just share a few examples. On the campus side, with the University of Texas system, they have integrated professional certificates, industry micro-credentials, for hundreds of thousands of students so far. More than 10,000 students have taken courses across 14 different campuses, more than 25,000 completions, and more and more of these campuses are integrating these professional certificates for credit into degree programs. In terms of Coursera for Business, one of our largest and best-known customers is L&T. This is a large construction conglomerate in India, and they've been using Coursera for years to train people in cybersecurity, software engineering, data analysis, and other things. But in particular, they have really gone aggressively to train people on generative AI across 20 different business segments. You might be wondering why is a construction company teaching people about Gen AI? Gen AI is for anyone who thinks. It is a way of thinking differently. It's a tool that could be used to enhance thinking. Anyone who decides, if you give it the right context, it can help with decision making. Anyone who communicates, if you take your point of view and you show it to it, or provide the context of what audience you're speaking with, it can really help you articulate your thinking much more effectively. So Gen AI really will change every job, assuming that people think, make decisions, and communicate ideas. L&T is the kind of company that realized this. I'll also say, last week, I can't name the name, one of the largest law firms in the world bought Gen AI Academy for every single lawyer, every single paralegal, every single staff member, because they know that fundamentally the nature of law will change: the way we practice it, the way we train for it, and the way that we work together where law firms work together to serve their clients. On the Coursera for Government side, we've been working with the State of New York. This is one of my favorite case studies because the Department of Labor has been collaborating with the Department of Education. The Department of Labor is offering professional certificates to underemployed and unemployed New York citizens. These certificates can create a pathway to a job, but they also could be a pathway to a degree. The State University of New York system, their Empire Campus which has 140 online degree programs, recognizes credit for the certificates being offered in the Department of Labor's workforce development program. So now you see collaboration among state agencies. We're also seeing the same thing at the national level, where ministries of labor and ministries of education are working together to create pathways to economic opportunity and pathways to learning opportunity. I'm also happy to say that the way that we're working to serve underserved populations through nonprofits in our social impact program has been growing as well. Today we work with more than a hundred nonprofits to serve learners who otherwise might not have access to this kind of material. I want to thank our partners, by the way. They make their content and credentials free to work with us and these nonprofits in serving these populations. Together, we've served more than 300,000 people, almost 2 million course enrollments in our social impact work. Relatedly, I was in Guyana recently with the president. The country of Guyana, led by the president, has basically bought Coursera and all the content and credentials on the Coursera platform for virtually every ministry and virtually every citizen. In the speech that he gave, there were about 800 nurses and students there because we had this big event where we launched this partnership. He said, 'It's okay to dream, but your dream can remain unrealized if the system does not support you in achieving it.' A lot of what we've been talking about here is not just individuals going to Coursera.org and taking courses. People need more support than that. So much of human activity exists within institutions, is supported by institutions. It is a privilege for us to work with the institutions that we have in this ecosystem, in this community that we call Coursera. So this is a phrase that is well known: talent is equally distributed, but opportunity is not. We are sitting at a time where the pace of change is accelerating dramatically, and unless we can have a rate of learning that matches that pace, this will become more true. But if we can use these tools and work together to figure out how to accelerate not only opportunity for learning but the rate of learning, we might reverse this. Hopefully what we've been able to show you today is the way that we've been working with this community to do a better job making better learning available to more people around the world in a way that provides pathways to opportunity. We're doing that together. It's a real privilege to have you here. It's an honor to work with you. Together, I feel very optimistic that we're going to make a difference in helping move humanity forward through education. Thank you.