By Christiane Martinez and Jenna Planden
There are many claims about the capabilities of technology, especially as it continues to influence education and other areas of our lives. Most of these claims emphasize the use of technology while ignoring the people, expertise, and techniques that go into its application. The media debate between Clark (1994) and Kozma (1994) revolves around this tension, inviting us to look at how the use of technology in education is discussed today. The following examples from Garmin (2022) and Duolingo (2023) demonstrate that this debate continues to be relevant.
Garmin (2022) reported a nerve-wracking flight captained by pilot Ryan Cianciolo. Despite facing drastic changes in cabin pressure, the plane landed safely before an emergency had to be declared. Garmin (2022) foregrounded Cianciolo’s assertion that this was all, “…thanks to the watch” (para. 2), which alerted him to changing air pressure. While not necessarily an example of educational technology, this admission is rife with techno-deterministic thinking and serves as an obvious illustration of the great media debate, along with some clever marketing from Garmin.
At its core, this debate is about whether media simply channels the methods of learning (Clark, 1994), or if the capabilities of media have the capacity to shape learning methods (Kozma, 1994). Clark (1994) would take issue with the way that Garmin (2022) presented the situation. Clark (1994) would see that the situation was resolved without catastrophe due to Cianciolo’s decision to set the watch, his previous education and training, knowledge of appropriate cabin pressure, and quick thinking to notify air traffic control (ATC). ATC could also take credit for their role in allowing Cianciolo’s rapid descent. The watch was simply an example of a “vehicle” (Clark, 1983, as cited in Clark, 1994, p. 22) for the actions of Cianciolo. Kozma (1994), on the other hand, would credit the technical capacities of the watch, together with Cianciolo’s training and personal attributes, and ATC, as the frame that prevented a serious emergency.
Now in the arena of educational technology, this kind of techno-deterministic thinking is demonstrated in Duolingo’s (2023) announcement of Duolingo Max, an AI-powered subscription tier built using GPT-4. Duolingo describes Max as its “most advanced learning experience” (para. 2) and suggests that its AI-powered features provide learners with the tools needed to advance their language skills. Two features are emphasized: Explain My Answer, which provides personalized explanations of learners’ mistakes, and Roleplay, which allows learners to practise conversations with AI characters and receive feedback. The company makes an even broader claim about the potential of the technology, with Duolingo CEO Luis von Ahn stating that AI will “shape the future of education” (para. 4) and may eventually recreate the experience of a human tutor at a global scale (Duolingo, 2023).
Clark (1994) would probably question the assumption that AI itself is responsible for improved learning. From his perspective, Duolingo Max serves as a vehicle for delivering instructional methods, such as practice, feedback, explanation, and simulated conversations. These methods, rather than GPT-4 itself, would account for any learning gains. Since a human tutor or another medium could potentially employ similar instructional strategies, Clark (1994) would therefore encourage us to separate the novelty and capabilities of generative AI from the instructional methods integrated into the application before concluding that AI is transforming learning.
Kozma (1994), however, would approach Duolingo Max differently. Rather than separating the medium from the instructional method, he would ask how the capabilities of AI interact with learners and the learning process. Kozma (1994) argues that media can contribute to learning when their distinctive capabilities are intentionally connected with cognitive and social processes. From this perspective, Duolingo Max’s ability to generate context-specific explanations, respond dynamically to learners, and create interactive conversational scenarios could matter because these capabilities determine how learners engage with and practice a language.
Garmin (2022) and Duolingo (2023) make strong claims about what technology can accomplish, while Clark (1994) and Kozma (1994) encourage us to consider the other factors involved. Clark (1994) highlights the methods, knowledge, and human actions that contribute to an outcome, while Kozma (1994) encourages us to consider how the unique capabilities of a medium can interact with these factors. As technologies such as generative AI become more widely used in education, the media debate helps us critically examine claims about the impact of technology and consider the factors that influence whether it actually supports learning.
References
Clark, R. E. (1994). Media will never influence learning. Educational Technology Research and Development, 42(2), 21–29. https://doi.org/10.1007/BF02299088
Duolingo. (2023, March 14). “Duolingo Max” shows the future of AI education. Duolingo, Inc. https://investors.duolingo.com/news-releases/news-release-details/duolingo-max-shows-future-ai-education
Garmin. (2022, September 29). Garmin smartwatch alerts pilot to decreasing cabin pressure. Garmin Blog. https://www.garmin.com/en-US/blog/aviation/garmin-smartwatch-alerts-pilot-to-climbing-cabin-pressure/
Kozma, R. B. (1994). Will media influence learning? Reframing the debate. Educational Technology Research and Development, 42(2), 7–19. https://doi.org/10.1007/BF02299087