Classrooms have become a battleground in the media frenzy around new large language models and other developments in artificial intelligence.
On the one hand, proponents of the technology jump to tout its promise for students. In each of its recent product launches, artificial intelligence forerunner OpenAI has partnered with marquee ed-tech companies and organizations to cast a feel-good halo around a technology that strikes many as apocalyptic.
One of those partners is Khan Academy, known for its zero-cost-to-consumer educational content. Sal Khan, its founder and global frontman, recently published, “Brave New Words,” which outlines a positive vision for how AI will uplift the world’s learners. One wonders, though, if he could have used a lesson on Aldous Huxley’s work before alluding to it in his own book title.
Bill Gates, a former software CEO who is now one of the world’s richest philanthropists, dedicates an entire section of his personal blog to how technology — chiefly AI — will usher in the future of education.
On the other hand, crafters of educational policy and educators themselves are pushing back — not just against AI, but increasingly against all digital technology in the classroom.
The Los Angeles Unified School District has become the largest district in the country to ban smartphones in classrooms. In New York, Gov. Kathy Hochul (D) is deliberating a similar ban for the state, while New York City Schools Chancellor David Banks has all but promised the city will follow the example of Los Angeles.
Sweden, once one of the world’s most aggressive nations in the digitization of its education system, has reversed course, among other things completely banning the use of digital educational tools in classrooms for students under six years old.
If the debate is over solutions, then both sides must ask: What’s the problem? If students are failing to meet expectations, why?
For the ostensibly progressive champions of information technology in the classroom, the problem is access. Access to information, access to the highest quality instructional materials. In the case of AI-powered tutors, access to one-on-one instruction.
For the regressive-seeming critics looking to de-screen the classroom, the problem is attention. Computers, particularly smartphones, shatter focus. Without focus, students cannot sustain the intellectual work required for learning.
Let’s take these problem statements one at a time.
On the problem of access, while the adoption of AI is too recent for a robust analysis of correlation with student outcomes, we can look at the educational results of other advancements in information technology to see what effects, if any, may be associated with their adoption.
Let’s start with the internet itself. According to the National Center for Education Statistics, a program under the United States Department of Education, internet connectivity among American public schools reached 94 percent in 2005.
So how did student outcomes change over the course of internet adoption in schools? According to the Program for International Student Assessment’s scores, student outcomes declined in every field:
- Reading: U.S. students’ scores decreased from 504 in 2000 to 495 in 2003, and then to 489 in 2006.
- Mathematics: Scores declined from 493 in 2000 to 483 in 2003, and then to 474 in 2006.
- Science: Scores decreased from 499 in 2000 to 491 in 2003, and then to 489 in 2006.
Now let’s turn to the smartphone. The iPhone launched in 2007; by 2015, most adolescents owned a smartphone or had access to one. At this point, not only did students have access to essentially unlimited information in principle, but they had that access with them at all times.
How did student outcomes change over those years? Looking again at the Program for International Student Assessment’s scores, this time from 2009 to 2015, we see flat or slightly declining scores:
- Reading: Scores declined from 500 in 2009 to 497 in 2015.
- Mathematics: Scores declined from 487 in 2009 to 470 in 2015.
- Science: Scores declined from 502 in 2009 to 496 in 2015.
Notably, the PISA score is normed to a global average, while the U.S. was one of the first countries whose schools adopted the internet.
The same is true for the smartphone. If either the internet or the smartphone benefited educational outcomes, we should see American students outpace their global peers. Instead, we see the opposite.
Now let’s turn to the problem implied by those aiming to remove technology from the classroom: attention. If the trends in information access have been positive, the trends in attention have been anything but.
Putting aside the intuitive causal relationship between these trends, there is ample evidence that students’ ability to sustain attention is declining.
According to the Centers for Disease Control and Prevention, the proportion of children and adolescents with an ADHD diagnosis nearly doubled between 1997 and 2016, rising from 6.9 percent to 10.2 percent.
Apart from the trends in diagnosis for attention-related disorders, there’s ample evidence to support the common sense notion that dividing attention between class content and one’s device produces negative effects on learning.
In one study following college students, texting during class was associated with lower GPA and test scores, no matter the amount. A similar study found that while immediate comprehension was not affected, long term retention was.
Other research has shown the mere presence of a phone can distract students and negatively impact their learning.
A 2017 study from the University of Chicago tested three scenarios in which the participants’ phones were present but not involved in or attended to during a learning task. In one setting, the phone was on the participants’ desks. In another, it was in a pocket or a bag. In the third, it was in another room.
The findings? “Paired comparisons revealed that participants in the “other room” condition performed significantly better on the [the cognitive task] than did those in the “desk” condition,” write the research authors.
Technology is a tool, a solution to a problem. The case for AI in the classroom needs to begin with an understanding of technology in the classroom, and an understanding of technology in the classroom needs to begin with an understanding of the problem to be solved.
The positive case for technology, that of access to information, can summon no evidence in its support despite decades of opportunity. The negative case, of shattered attention, is tragically gaining in evidence, not just as an assessment of the minds of our students, but also in negative effects on educational outcomes.
In his last taped lecture, recorded in 1979, media theorist Marshall McLuhan popularized an adage from Father John Culkin: “We become what we behold.”
“We shape our tools, and thereafter our tools shape us,” Culkin continued.
Our students are being shaped by smartphones and other devices; that shape may be less smart than we hope for them. And we’d be wise to take our lessons there, lest artificial intelligence shape them to be artificially intelligent.
Thomas Howell is the founder of Forum Education.