daily dose of aspiration #10

Chegg and the rise of AI tutors, methodology > content, futurist classroom

Hi! Welcome to my daily blog series. This series consists of one company in the Edtech space that I discover, one learning I have from that day, one idea, and one aspiration I have as I experience life post-grad.

🛠 One Company

Today I will invite you to learn about Chegg, an Edtech unicorn serving as one of the most widely used homework help solutions in the world. However, the birth of ChatGPT has significantly impacted the customer growth rate of Chegg.

Chegg’s CEO said:

Since March (2023) we saw a significant spike in student interest in ChatGPT. We now believe it’s having an impact on our new customer growth rate.

Chegg’s quarterly revenue was down 7% year-over-year (down to $187.6 million) as it became the first publicly-traded company to admit the direct negative impact ChatGPT had on its product.

Just prior to the rise of ChatGPT, Chegg’s annual revenues and stock prices skyrocketed during the COVID pandemic in 2020 and 2021. ChatGPT has clearly shifted the growth trajectory of Chegg.

Chegg’s annual revenues growth prior to, during, and post-COVID pandemic.

Chegg started as a book rental company before providing more services such as studying and homework help. The company at the moment has three lines of products:

Book

Chegg provides students with the option to rent/buy textbooks at a lower cost, read online, or sell books.

Study

Chegg provides students with endless tools and services to check answers to their homework, support and enhance writing, and prepare for exams through study tools such as practice problems, practice exams, and flashcards.

Career

Chegg provides online courses and internships for students to practice skills and gain internship experiences necessary for their career development.

To add to its Career services, in September 2019, Chegg acquired the Online Skills-Based Learning Platform Thinkful to help students accelerate their path from learning to earning.

What does it mean ‘from learning to earning’? Chegg and Thinkful provide students with the opportunity to learn first, get hired, and pay tuition later. The platform provides tracks mainly in tech such as data science, software engineering, and UX/UI design. Students have the options to either study in a cohort or study independently. If students don’t get hired in a qualifying position within 6 months of graduating, they get a full tuition refund.

The programs that Chegg currently has include: software engineering, data analytics, UX/UI design, data science, technical project management, and digital marketing.

The student outcome the company focuses on is the hire rate and increase in average salary to showcase Chegg’s impact on helping students move from learning to earning to kickstart their careers.

Interestingly enough, after the decrease in its customer growth caused by ChatGPT, Chegg has recently launched another line of products called ‘CheggMate’ which is partnered with OpenAI, the company that founded ChatGPT. CheggMate is a GPT-4 powered AI learning assistant that combines the power of GPT-4’s advanced AI systems with Chegg's extensive content library and subject-matter experts to make learning more:

• accurate

• adaptive

• faster

• personalized

The unique selling point of CheggMate as compared to other AI tutors in the market might be Chegg’s proprietary data of billions of content from many textbooks Chegg has accumulated over the years. As a student trying to get help on a specific topic in school, CheggMate might provide me with more accurate answers than ChatGPT. In this case, CheggMate targets students specifically as compared to serving any human being which is what ChatGPT is doing. I think that by building CheggMate, Chegg is trying to win back students around the world by providing a more accurate homework solution than ChatGPT.

Waitlist for CheggMate is currently open to signing up.

🧠 One learning

So far, we have seen a rising trend in the intersection of AI and education: AI tutor! From Synthesis’s Digital Tutor, Khan Academy’s Khanamigo to Chegg’s CheggMate, an AI learning assistant for every student will be an increase in demand as more companies figure out a way to build this technology. Just a few months ago, I wrote about an application of AI in education - to make learning more personalized. I didn’t predict AI tutors to have been born already … A few learnings and thoughts I have from the rise of AI tutors:

• The need for more personalization: Similar to how Workera.AI is using AI to figure out granular skills that each employee should focus on, AI tutors for students from K-12 and higher education will find a way to make their products as tailored and specific to their audience as possible. An AI tutor who specialized in math, an AI tutor who knows all the textbooks in biology, an AI tutor who can solve college-level economics, etc.

• The battle for content accuracy (is no longer valid 🧐): Among the current AI tutors, one of CheggMate’s unique features is being trained by subject matter experts who quality control the accuracy of the content. I would argue that the accuracy of content is the wrong premise to start with because as AI becomes more developed and better trained, is the accuracy of the content that important anymore? If we know this content or that content is accurate, so what? Will subject matter experts’ intelligence be more valued than AI intelligence? Should subject matter experts keep battling with AI? This leads to my last thought.

• The importance of methodologies/pedagogies in learning and teaching: As AI tutors become more possible and effective, the most effective online learning products won’t be those that give you the most accurate answer (the best content) but those who guide you through your learning struggles most effectively (the best methodology to learning) (also shared by Alberto from Transcend in his recent post). Learning is not about getting the right answer and proving that you are right all the time. Learning is about processing, thinking, struggling, and reflecting on what one has learned.

We have passed the era of getting information onto the Internet, Edtech products need to focus on ways to help people learn the information that is widely available on the Internet.

Content-first companies will be easily replaced by AI. Methodology-first companies won’t.

💡One idea

I think school will be a much better place if we can allow students to say they don’t know everything yet and if they can collaborate healthily to achieve collective intelligence.

I recommend this post by my housemate where she draws a vision of a future classroom where students can together say we don’t know shit (yet!).

I imagine a future classroom where I can comfortably tell people that I’m dumb and people will say “Me too” and laugh together. I imagine a future society where I say I don’t know shit to someone without feeling insecure about my first impression. I want to learn and thrive in a safe playground and celebrate I don’t know shit and people don’t know shit either.

🤩 One aspiration

I want to create a learning hub where students feel safe saying that they don’t know and where students collaborate with each other and with AI tools to co-create knowledge together.