Jeff Maggioncalda, Coursera CEO: A Fortt Knox Update

Well, there's a lot that we're doing with Coursera Coach. The start of Coursera Coach was helping learners learn more effectively by basically providing a tutor that can help them understand and practice concepts in the course. But what we announced today, and a number of professors from universities are talking about it, is Coach as an extension of the teacher. So basically, the authors—this is Vic Strecher from the University of Michigan who was on stage today—he's using Coach, he's programming Coach, which is an artificial intelligent agent, to act as an extension of him in providing personalized interactive instruction in the Coursera courses. So it's no longer just videos; you're actually interacting with the instructor via Coach. So that's one of the big things that we did with Coach.

Well, this is what's really neat about it. Even when Coach was acting as a tutor, basically when a learner asks a question of Coach, Coach reads through the whole professor's course and finds the right answers so that the answers are not coming from the internet; they're coming from inside the course. What the professor is able to do now is use Coach as an extension of themselves as a teacher. So basically, what the instructor does is they say to Coach, 'First of all, you're in my course and I want you to really stay grounded in my course content.' Then they actually tell it instructions: 'Hey Coach, I want you to teach my students A, B, C, and D. Here's what mastery looks like, and here's the style and tone that I want you to use. But keep your answers and keep this dialogue grounded in my course. Don't go off and find information someplace else. I want to be instructing the students based on the material in my course.'

Well, this is what's also really cool. You know, universities and industry leaders like IBM, Microsoft, and Google really are careful when they build their courses. Since their brand is on their courses, they make sure that the intellectual property is appropriately used. And so, to the extent that over the last 12 years, 7,000 courses have been built and there's an author behind it who's responsible for it, Coach is just using the content that's in the course. Now, one other thing about the Coach which is kind of cool: as an instructor, the instructor can also load in their lectures, any material that they've written, as resources for Coach to do the instruction. But it all happens based on the information that the instructor has put in. And so, the instructor is responsible for making sure that the rights to use that material are there, and that's been true for over a decade on Coursera. They've always respected the intellectual property rights.

There's always been a lot of feedback in the courses. For the last decade before generative AI, students could give thumbs up and thumbs down on any item in a course—a video or a reading or a quiz. They can also directly message the instructor and say, 'Hey, I like this or I don't like this.' With respect to Coach, we record the conversations so that the instructors can actually see, not only discrete thumbs up or thumbs down, but they can actually see the way that learners are interacting in these dialogue sessions. And we don't yet do this, but you can easily imagine a case where we could summarize some of these interactions so the instructors get a clearer sense more quickly of what are the challenges and the opportunities to improve their courses and the effectiveness of the learning.

Yeah, we'll start with the demand. We are seeing a lot of demand for generative AI content. We're getting six enrollments per minute in generative AI content. Our partners have produced more than 500 generative AI titles. We now have about three million enrollments that have happened in generative AI content. And we just announced today a lot more content. Now we have something called Generative AI Academy. This is a collection of courses and projects and other skills and tools to help people learn about this. We started with Generative AI for Everyone, and that's a set of courses that are general purpose: you know, how does it work, how do you use it responsibly. We have a set of courses called Generative AI for Executives—that's how do you navigate your institution through it. And then, recently in the last couple months, we announced Generative AI for Teams, and this is sort of the big one because a lot of companies are looking to unlock productivity gains. McKinsey thinks there's $4 trillion of productivity gains. But to do that, you really need to teach software engineers how to use GenAI to do software engineering, or cybersecurity, or data analysis, or product management. So we've just announced today a set of titles from Microsoft and IBM for business intelligence, cybersecurity, data analyst, data engineering, data science, project managers, product managers, software developers. All these job-specific training programs. And then Google announced today that they have upgraded all of their entry-level career certificates, which are the most popular titles on all of Coursera, to include generative AI modules to teach people how to do these jobs using the new tools of generative AI.

You know, I thought that when we announced all the job-specific titles, that's where most of the attention would go. As it turns out, right now a lot of people are still in stage one: what is this stuff? What is generative AI? What does it mean? How does it work? More importantly, how is it used? I think the next wave, what we're starting to see, is people saying, 'Okay, now that I kind of know what GenAI is'—these are courses like Generative AI for Everyone, Generative AI for being a thought partner, how to use it responsibly—but now, as people get their arms around the concept, they're thinking a lot more about how do I do my job differently. So they're going to be waves of this. And I'd say right now, the interest still is pretty general purpose. But as we put more of these job-specific training programs and certification programs in place, I think that's where the ball's going next.

Here's what's pretty interesting. We serve individuals and we also serve institutions. On the individual side, people coming to Coursera from emerging markets like India, Latin America, Southeast Asia—when we survey learners, they're much more optimistic and positive about the impact of generative AI on job opportunities in those countries. There's some skepticism in the US. And what I think is going to be pretty interesting is that company after company is saying that they would rather hire people who have these skills and can be more productive, whether they're in the US or someone else in the world. So I think there's a little bit of an opportunity and a threat here. In emerging labor markets, we're seeing a lot of demand from learners to learn these skills. I think in some of the more developed, wealthier nations, there's a little bit more worry. But as you said, if someone doesn't learn the skills, it might not be AI that replaces your job—it's somebody else in the world who's learned how to use AI that's going to replace you in your job. And so individuals are starting to warm up to that. On the business side, from law firms to construction companies, businesses are realizing that anyone who thinks, anyone who makes decisions, anyone who communicates ideas will benefit from a thinking tool called generative AI. So we're seeing very broad interest among businesses for these kinds of titles.

You know, we're definitely seeing that correlation between labor markets and educational opportunities. We're seeing good uptake among universities. I mean, one of the things that's happening is that universities are starting to realize that generative AI not only changes what employers are looking for from the graduates, it also changes what students are looking for from the course offerings, it also changes how they learn, and it also changes how the teachers teach. So honestly, universities are completely ripe for major change because of this. And they're starting to embrace these industry microcredentials. They're integrating these industry microcredentials into their degree programs in order to keep up with how fast things are changing. I mean, no one can really do it themselves. Collaboration is kind of the theme here to keep up with the rate of change.

Well, you know, it's interesting. I was really struck when I first used GPT in December of 2022. We immediately jumped on this. And it's interesting because I created a course called Navigating Generative AI: A CEO Playbook. A couple professors at Harvard were taking my course. They said, 'Hey, this is a pretty interesting example of how a company is responding to it.' So they created a Harvard Business School case about Coursera's response to generative AI. You can take my course or someone could take the Harvard Business School case, which is now being taught all around the world. I guess what I would say is we responded really fast. We realized things were changing. We can't really predict the future, but we knew if we didn't learn quickly about the threats and opportunities of this, it would certainly become a threat. And so we're just trying to move as fast as we can, to learn as fast as we can with our partners, to create valuable things for customers. Like that's what this is about: how can we solve problems and create value that's relevant for our customers using this new technology?