On a return flight from Tokyo to Detroit, I impulsively downloaded the Duolingo app. As a professor of education practice, I had just attended an educational technology conference and was inspired by how colleagues were using the app to teach themselves a new language like French or Spanish. Maybe Duolingo could help me learn Japanese and prepare for a return trip my husband and I were planning.
As we prepared for takeoff, I noticed how the app established credibility by stating its teaching approach was rooted in educational research. It used techniques like “spaced repetition” to help people learn. As I worked through the introductory exercises, I recognized elements of “gameful learning” such as featuring stars and leaderboards. These would make useful examples within courses I teach about games and learning. The app’s interface was slick and motivating. I was drawn in by its quirky avatars and my ability to earn quick wins in a variety of modalities like text, speech and audio. As I settled into my long flight, I felt confident that I could learn a new language.
The following week, Duolingo made headlines when its CEO, Luis von Ahn, declared that AI is a better teacher than humans and that schools will only exist “because you still need childcare.” This viewpoint was reflected in the company’s “AI-first” strategy, which replaced workers with artificial intelligence tools.
The backlash from Duolingo users was swift. Outraged, many of my learning sciences and educational technology colleagues in academia and industry posted on social media that they were deleting the app even after years of dedicated use, commenting that they had noticed a downturn in the quality of Duolingo lessons anyway. When the CEO of an ed-tech company valued at $23 billion can casually dismiss the entire education profession as glorified childcare, we should all be alarmed.
The Duolingo moment perfectly embodies something that I’ve noticed recently—that the ed-tech landscape is experiencing a dangerous shift away from educator expertise toward an AI-driven, technology-first approach to digital innovation, ignoring designs for learning that are based on what we know about how people learn. And although Duolingo has since tried to backtrack on its “AI-first” strategy and even distance itself from its CEO, this event sheds light on a growing conflict.
There is an uneasy alliance between educators and ed-tech venture capitalists, which allows both groups to achieve certain objectives. For educators, access to “out-of-the-box” educational tools means they can be used in educational settings without a lot of customization and configuration. This can be especially beneficial with the advancement of AI-driven tools that require significant funding to develop and maintain. For ed-tech venture capitalists, access to innovative use cases and educator testimonials can help their bottom line. And if ed-tech companies can boast that their tools are based on educational research and thus be purported to “work,” even better.
Von Ahn’s remarks are indicative of a larger trend. Big Tech approaches education like any other market—to disrupt. Too often, however, they do so without sufficient regard for educator expertise and experience. They are building solutions for learning problems that they do not fully understand, replacing research-based insights with algorithm-driven designs.
Recently OpenAI announced its goal of making ChatGPT an integrated part of every college student’s campus experience. With a personalized account, newly matriculated students will have access to AI tools built into learning management systems that support them “from orientation day through graduation.” Student data obtained from AI learning tools embedded in their courses will be used to power AI chatbots that will provide students with a personalized college experience through coaching, tutoring and career services.
This wholesale and mandatory adoption of powerful AI tools should be approached with great caution, particularly as new research indicates that AI use in education could erode critical thinking skills and bypass the acquisition of foundational domain knowledge.
To be sure, Big Tech brings educational tools to market that educational and research enterprises could not realize on their own, meeting demand for self-directed learning opportunities and providing tools that will help students prepare for a changing workforce that includes AI. But to fully realize these opportunities, we need research-based solutions, not venture capital experiments that are no more effective than throwing spaghetti at the wall to see what sticks.
We can either accept this shift toward big industry dominance and its disregard for expertise and classroom experience, or we can demand that educational tools be built in true partnership with teachers and educational researchers.
A recent Cornell report illuminates the need for educators to be included in AI tool development and implementation. The report recommends that technologists focus on solutions to technical challenges, such as preventing AI hallucinations and addressing privacy concerns, while educators focus on broader issues, such as the potential for AI tools to inhibit critical thinking skills and social development. By bringing the perspectives of technologists and educators together, we can create tools that better support student learning and minimize harm to educational ecosystems.
The Christensen Institute proposes ways to foster productive partnerships between educators and technologists, such as by businesses embracing models that prioritize ed-tech partnerships, by fostering collaboration and shared language, by using research-based teaching practices to guide technology development, and by testing design assumptions in authentic educational settings. The Cornell report offers additional guidance, including prioritizing educator input regarding whether and how to implement AI tools and designing tools that give educators agency to customize them and question and correct AI outputs.
If we embrace a partnership model, educators and ethical ed-tech companies can enter productive alliances with mutual regard and respect for each other’s expertise and strengths. The choice we make now will impact the quality and effectiveness of learning for generations to come.
