April 10, 2024

GPTd on Arrival: A New Paradigm for Understanding Education Technologies

Categories: Connected Learning, Critical Perspectives, Digital Learning, Edtech, Educational Practice, Featured, Research

Image: Generated in DallE

Over the past 125 years, since the advent of applications for radio and film strips in schools – and on through television, terminal computers, personal computers, smartboards, tablets, laptops and other devices – the dominant model of technology integration in schools has been adoption. School administrators, policymakers, and other education leaders deliberate about the potential benefits of a particular technology, conduct pilot tests, allocate funds for purchase, initiate procurement, develop usage policies, provide pedagogical training for teachers, launch classroom implementation, and track usage, impact, and outcomes. Of course, not every adoption process involves all these steps, but historically, education technology integration has involved some degree of strategy and planning.

Generative AI was not adopted; it arrived. 

In contrast to adopted technologies, “arrival technologies” bypass the planning, assessment, policy-making, and professional learning that have historically (if imperfectly) accompanied previous generations of technology integration. Some prior innovations could be classified, at least partially, as arrival technologies – students brought their personal calculators to math class in the 20th century; mobile phones brought the internet into some classrooms before intentional adoption – but generative AI represents a step change in both the velocity and nature of technology arrival. In less than a year, nearly every internet-connected computing device suddenly had access to dramatic new capacities. Moreover, generative AI is likely to arrive in schools not only on student and teacher devices, but in a wide array of existing software in schools. Student information systems, learning management systems, collaborative writing tools, plagiarism detection software, intelligent tutors, and other software have been integrated into schools through traditional adoption processes, and in the months and years ahead, new AI capacities will arrive in these systems without the same planning, intention, and oversight.


“…generative AI represents a step change in both the velocity and nature of technology arrival.”

Substantial portions of the corpus of education technology research and policy guidance rests on the adoption assumption. In the recently published National Education Technology Plan, for example, the word “adoption” appears 30 times. Here are two instances:

  • “Develop rubrics for digital resource and technology adoptions to ensure tools are accessible and integrated into the larger educational ecosystem, support Universal Design for Learning (UDL) principles, and can be customized in response to accommodation or modification needs of learners with disabilities.”
  • “With finite time and funding, it is incumbent upon education systems to verify the effectiveness of technological tools before purchase and adoption, and during classroom implementation.”

For educational systems to effectively address generative AI, and for education research to study this new integration pathway, we need new theories of education technology that account for arrival technologies.

For instance, arrival technologies might have very different implications for education equity than adopted technologies. Educational technologies can cause harms to educational systems as well as benefits: distraction, academic misconduct, reduced instructional quality, and cyberbullying are all real risks of education technology. With technology adoption, schools can introduce new technologies when they are prepared to manage these risks; and they can make the important, prudent choice not to adopt technologies. Arrival technologies impose harms and benefits on schools whether or not schools choose to allocate resources to their integration. ChatGPT arrived on the shores of U.S. schools at a moment where well-resourced schools in affluent neighbors have made significant progress in recovering from the pandemic, and where chronic absenteeism, teacher shortages, and unfinished learning remain all-consuming challenges for under-resourced schools. Our initial research suggests that well-resourced schools can flexibly devote staff resources to assessing new AI technologies, developing policy, and addressing potential sources of harm. Under-resourced schools are forced to deal with the consequences of AI arrival without the adequate resources.

A robust theory on arrival technologies will aid policymakers, administrators, and educators in managing the arrival of generative AI and future generations of education technology that bypass adoption.

Guest post by Justin Reich

Justin Reich is an associate professor at MIT, and the author of Iterate: The Secret to Innovation in Schools. He recently co-authored some insights and recommendations on GenAI in schools: Generative AI and K-12 Education: An MIT Perspective.