Artificial but Intelligent Education

Our system of public primary and secondary education is, generally, in deescalating shambles. Budget cuts – ranging from attempting to apply unworkable business tax cut incentives (e.g., bringing the Kansas public educational system to the precipice of collapse under another example of the failed “trickle down” economics theory) to sapping by reason of the pandemic to a simple effort to downsize – have combined with reallocating resources to follow politically popular but anti-academic programs of instruction (e.g., creationism, perceived censorship of historical facts to placate political pressures, etc.). The result is the plummet, applying international standards used to evaluate educational skills at varying levels, of U.S. high school performance from first to nineteenth to thirty-eighth depending on subject area.

We are finding a powerful dearth of STEM-competent workers, a failing that starts with our secondary preparation needed to qualify and complete a relevant mathematically based college/university degree. We have had to reach to other nations for the relevant expertise, with tons of high-level jobs unfilled for lack of local candidates. As immigration rules were tightened, even as the pandemic is ending, too many of those jobs remain simply unfilled. Too many students. Deferred maintenance schools and classrooms. Budget cuts. 

Classroom size varies across the United States, but there were no states with top ten cities with fewer than 25 students per average public high school class, most over 25 in any event. California has a staggeringly poor 35 such students per class. Teachers distracted with fixed and often educationally irrelevant curricula, many with outmoded textbooks and ill-equipped classroom facilities, are forced to prepare students for a demanding globally competitive labor market. The slam to the educational process brought on by the pandemic, where many students were simply unable to adapt to remote learning for any number of reasons (lack of computers, lack of internet access, discomfort with the technology, distracting home environments, lack of discipline, etc.), was somewhat countered by the distribution of computers to students who might otherwise not be able to afford them.

If we expect to prepare the rising generations of students for careers in the future, they are going to have to adapt to increasing automation and artificially intelligent data processing and analytics. Training coal miners needs to be replaced with training those who can install and maintain alternative energy systems, for example. Does AI offer any possible contributions to upgrade our public educational systems? The obvious answer is yes, but then there is huge question: who is determining what information is being fed into the overall system being provided to all relevant students? Think of the ability of the information provider to shape our entire society. Smelling any ethical issues? Google seems to have developed the most likely platform for the next (current?) stage of public primary and secondary education in at least the United States.

Here’s the way Ben Williamson, a senior research fellow at the University of Edinburgh, put it in the May 28th FastCompany.com: “When Google CEO Sundar Pichai addressed the company’s annual I/O Developers Conference on May 18, 2021, he made two announcements suggesting Google is now the world’s most powerful organization in education. Opening the livestreamed keynote from the Mountain View campus gardens, Pichai celebrated how Google had been able to ‘help students and teachers continue learning from anywhere’ during the pandemic.

“Minutes later, he announced Google’s new AI language platform, a central part of the company’s long-term AI strategy, with a specific use-case example from education. LaMDA (Language Model for Dialogue Applications), he claimed, could enable students to ask natural language questions and receive sensible, factual, and interesting conversational responses.

“‘So if a student wanted to discover more about space,’ Pichai wrote on the company blog, ‘the model would give sensible responses, making learning even more fun and engaging. If that student then wanted to switch over to a different topic,’ he added, ‘LaMDA could continue the conversation without any retraining.’ The company plan is to embed LaMDA in its Workspace suite of cloud computing tools, software, and products.

These proclamations indicate how Google plans to advance its business in education following the disruptions of COVID-19—by consolidating the huge growth of its platforms in schools and integrating AI into teaching and learning. That’s raising fresh concerns among privacy campaigners and researchers because it gives Google access to data about students and schools at international scale…  

With schools reopening worldwide, Google has worked hard to ensure the big market gains it made in 2020 can be sustained and strengthened as students return to physical rather than virtual classrooms. With user numbers of its digital learning platform, Google Classroom, up to 150 million from 40 million just a year before, it announced a new ‘road map’ for the platform in early 2021…

“Large language model technologies are among the most contentious of Google’s recent developments. Late last year, a group of researchers, including the two co-leads of the Ethical AI team at Google itself, produced a research paper claiming harmful ideas, biases, and misleading information are embedded in these models. Google subsequently fired the authors from its Ethical AI team, leading to widespread condemnation and serious questions about the long-term ethical implications of its AI strategy.

“This raises the troubling question of whether installing Google’s language AI technologies in educational products might reproduce biases and misinformation within the institutions of schooling. At I/O, Pichai maintained further development will ensure ‘fairness, accuracy, safety, and privacy’ are baked in to LaMDA before full rollout, though the firing of its Ethical AI specialists weakens the credibility of these assertions…

“Google has produced the hardware, software, and underlying cloud and data systems on which education systems are increasingly dependent, at scales that cross geographical and political borders and continents. These are technical, ethical, and political issues that should not only be delegated to educators and school leaders to sort out. They need to be addressed at the regulatory level, and through democratic, collective discussion about the future of schools beyond the pandemic.” And exactly who is going to make these inquiries, and what governmental entities are sufficiently staffed and qualified to contain the potential harm? Houston – and every other school district in the United States (there are about 13 thousand of them) – we have a problem? Beyond mere antitrust issues.

I’m Peter Dekom, and the least qualified societal American sector able to react to such hyper-accelerating change appears to be the government of the United States itself.