Home > Uncategorized > Elizabeth Mason Posits Three Ways AI Could Remove Poverty… One Looks Promising, All are Chilling and Unlikely

Elizabeth Mason Posits Three Ways AI Could Remove Poverty… One Looks Promising, All are Chilling and Unlikely

January 2, 2018

Yesterday’s NYTimes featured a thought provoking article on the promise of AI in addressing the intractable problem of poverty by Elisabeth A. Mason, the founding director of the Stanford Poverty and Technology Lab and a senior adviser at the Stanford Center on Poverty and Inequality. Ms. Mason posts three ways AI could help address the three underlying causes of poverty: joblessness; lack of education; and dependence on government programs. Of the three ideas Ms. Mason presents, I completely agree with her ideas about education but found her ideas about joblessness and welfare chilling.

First, Ms. Mason’s thoughts on how technology could assist education mirror mine, which, naturally, makes them appealing to me. She suggests that AI could be used to completely differentiate instruction, matching learning preferences to each child and pacing instruction in a fashion that enables all students to progress through the curriculum successfully. Here’s her description of how we currently treat students in K-12 schools and how AI might help:

We bundle students into a room, use the same method of instruction and hope for the best. A.I. can improve this state of affairs. Even within the context of a standardized curriculum, A.I. “tutors” can home in on and correct for each student’s weaknesses, adapt coursework to his or her learning style and keep the student engaged.

Today’s dominant type of A.I., also known as machine learning, permits computer programs to become more accurate — to learn, if you will — as they absorb data and correlate it with known examples from other data sets. In this way, the A.I. “tutor” becomes increasingly effective at matching a student’s needs as it spends more time seeing what works to improve performance.

I do NOT believe “machine learning” can replace teachers. But I DO believe “machine learning” can provide a better means for students to master hierarchical subjects like mathematics and basic science, rule-based topics like grammar, and skill development that is acquired through repetition. That, in turn, could free teachers to cultivate higher order thinking skills and develop interpersonal skills.

Ms. Mason’s idea for job-matching and welfare reform, though, are problematic. For job creation she asserts that AI is ideally suited for matching job seekers with job vacancies and further contends that AI can “…take the guesswork out of which jobs are available and which skills workers need to fill them.” As for welfare reform, Ms. Mason believes AI can “…predict which programs help certain people at a given time and to quickly assess whether programs are having the desired effect.

My problem with these algorithmic “solutions” is that they all require the collection and storing of massive amounts of data by a third party who would, presumably, use the data to achieve noble purposes. In the world as I know it now, I would expect our legislators to see the private sector as the ideal collector and manager of data, which would provide some corporation with a wealth of data that could be used to advance the unending consumption of resources so that corporations could achieve ever higher profits. Moreover, in their desire to achieve profits, I doubt that shareholders would place the public well-being over earnings. If that was the case currently, there would be no poverty or joblessness.

As an educator who has espoused individualization of instruction since I attended graduate school in the early 1970s, I despair at the idea of changing the existing paradigm of schooling where, as Ms. Mason describes, “we bundle students into a room, use the same method of instruction and hope for the best.” As long as we group students by age cohorts and use some form of standardized tests to measure their progress we will never achieve the kind of individualization Ms. Mason advocates.

Finally, as one who believes that AI could be used for beneficial ends, I despair even more that our current government dismisses evidence-based decision making… and that use of evidence-based decision making is at the heart of Ms. Mason’s conviction that AI has promise. Here’s her argument, presented near the end of her article:

Even Congress occasionally wakes up from its partisan slumber to advance the cause of technology in public policy decision-making: In 2016, Congress voted for and President Barack Obama authorized the creation of the Commission on Evidence-Based Policy Making. The act creating the commission was sponsored by Senator Patty Murray, a Democrat, and Paul Ryan, the House speaker. Before the commission expired in September 2017, it used government data to evaluate the effectiveness of government policy and made recommendations based on its findings.

This just in, Ms. Mason: the commission expired in September and none of its recommendations have been heeded, no laws have been introduced to implement any of the recommendations, and scores of evidence-based concepts have been abandoned by the current administration.

In the end, I do not see much promise in “evidence based” decision making, particularly if we maintain the toxic partisanship in our government and in our media coverage. In general, people can only absorb evidence that confirms their beliefs and validates their personal experiences. Until voters are willing to open their minds to the possibility that the narratives the political parties are promulgating are invalid, we will continue to muddle along basing our conclusions about “what works” on the convictions we hold.

%d bloggers like this: