With its publishing business in free fall, Pearson is targeting teaching, promoting Aida, its artificial intelligence program to teach calculus. Pearson claims Aida will provide a “personalized learning experience” and that the company is the first to apply “consumer apps” in education. Even if you don’t teach calculus, teachers globally need to worry about what Pearson aims to do about teaching as work. According to the Aida website, “By combining AI with the learning sciences – psychology, neuroscience, linguistics, sociology and anthropology – we gain an understanding of what and how people learn. With AI, how people learn will start to become very different.” AI advocates promote a science fiction future with everybody plugged in and learning at their own and at a higher rate. It is an enormous potential market for Pearson, replacing 3,000,000 American teachers, with its algorithms. 
Yet,  preliminary studies don’t support an AI learning revolution.  The Bill & Melinda Gates Foundation funded an Adaptive Learning Market Acceleration Program (ALMAP) to establish “evidence-based understanding of how adaptive learning technologies such as adaptive courseware could improve opportunities for low-income adults to learn and to complete postsecondary credentials.” It was of limited use because the study primarily looked at student and teacher attitudes about AI and largely ignored its impact on student learning and school success. The study’s Executive Summary apologized for AI because of the “relative immaturity of the field of adaptive learning technology.” However, it assured educators and investors that “technology capacity and ways to support instruction and learning” were “evolving rapidly.” A major concern they found with AI products was actually getting students to use the courseware.

Other findings were that AI worked better at a micro level (tutoring) than on a macro level (classroom instruction); that marketing claims for a variety of AI products were suspect; and that “multiple factors affect learning outcomes and to make sense of student outcomes, analyses need to incorporate student characteristics, specifics of how the adaptive courseware is used, aspects of the course beyond the courseware product, and the way learning is measured to make sense of student outcomes.” In other words, we know very little about the benefits of AI in education, other than while initial costs were high, long term costs, once you get rid of teachers, will be significantly lower.

Recent studies by Nobel Memorial Prize winning micro-economists confirm the drawback of schools making heavy investments in untested technology. One study, conducted in Kenya, found that just adding resources to schools improved educational performance by top students, but not everybody else, increasing social and educational inequality. Another study focused on schools in Mumbai, India where government officials are in love with new technologies. This study found that, across the board, students performed better when schools used extra money to hire additional teachers. The study’s authors found that “there exists very little rigorous evidence on the impact of computers on educational outcomes and no reliable evidence for India or other developing countries.”

The Balsakhi Program placed a second adult in a third or fourth grade classroom, usually a young woman from the local community who had finished secondary school. The Balsakhi worked with specific children who were falling behind academically, focusing on skills they should have mastered in earlier grades. While balsakhi had only two weeks of preparation prior to starting their positions, they received continuous support and instruction during the school year. The study found that all student subgroups improved their academic performance on standardized exams as a result of the Balsakhi Program and in some cases student scores were significantly higher. At the time of the study, the Balsakhi Program was supporting hundreds of thousands of children. The authors of the study believe a major reason for the success of the program is that the balsakhi were from the locality and better understood the academic and social needs of the children. For India, a side benefit of the Balsakhi Program and a reason it expanded rapidly is that it has “very low overhead and capital costs.”

And a UNESCO report questioned the ethical application of online AI instruction, given data breaches and the misuse of information to influence public opinion and to sell products.

But why should any of this stop Pearson from making money?

By Alan Singer, Hofstra University

You can follow Alan Singer on Twitter: https://twitter.com/ReecesPieces8