The learning curve for AI in schools

The learning curve for AI in schools

The rush to use artificial intelligence in education has led to insights on the types of intelligence – even new types – that students and teachers must bring to the AI era.

As American students head back to the classroom, their schools are being flooded with offers of AI-powered products to improve learning. This summer, Google and Anthropic both launched new technologies designed for teachers, joining similar offerings from OpenAI and Khan Academy. In addition, more than 30 states have adopted guidelines on the use of artificial intelligence in education.

A large majority of high schoolers, of course, already use AI for schoolwork. And last year, the percentage of teachers using AI tools almost doubled from the year before – from 32% to 61%.

This rush to adopt yet another new technology in classrooms will almost certainly help teachers with basic administrative tasks. At the same time, much of the research reveals that students will benefit when they bring their own special intelligence about the limits and dangers of AI, such as biased answers or a possible invasion of privacy.

About one-quarter of teens have asked a teacher how to judge the accuracy of an AI summary, according to a survey this year by the nonprofit Common Sense Media. Most teens see AI as both an answer machine and a way to support their thinking. Two-thirds say it helps them understand their school lessons, not just complete them more quickly.

The most profound result of the survey was that at least half of teens acknowledge the skills that AI can’t or shouldn’t replace: “working well with others” and “fact-checking and critical thinking.”

The new AI products for schools, in other words, are forcing new insights on intelligence itself, as well as which types of intelligence are best taught by teachers.

One leader in defining types of intelligence, Howard Gardner of Harvard University, predicted at a conference last year which types will probably remain anchored in human teaching.

“I think most cognitive aspects of mind – the disciplined mind, the synthesizing mind, and the creative mind – will be done so well by large language machines and mechanisms that whether we do them as humans will be optional,” said Dr. Gardner, a professor of cognition and education.

“On the other hand,” he added, “I don’t believe for a minute that aspects of respect – how we deal with other human beings – and ethics – how we deal with difficult issues as citizens, as professionals – can or should be consigned to even the most articulate and multifaceted, intelligent machines.”

Dr. Gardner told Education Week this year that AI might push humans toward an expression of a higher intelligence. “Perhaps we’ll even discover or create new knowledge, entities, or experiences beyond AI’s reach – maybe even poetry or stories no AI could write.” But, he continued, “that will only happen if individuals are intrinsically motivated to stretch their abilities.”

For schools, that is the rub: whether AI brings a unity and elevated understanding among students. “What matters,” Dr. Gardner concluded, “is that young people learn to work together, particularly on challenging ethical and moral issues.”