What Past Education Tech Failures Can Teach Us About the Future of AI in Schools

This article was originally published ConversationThe

American technicians are telling teachers to quickly take their new discoveries for more than a century. In 1922, Thomas Edison announced In the near future, all school textbooks will be replaced by film strips, as the text was 2% skilled, but the film was 100% efficient. These bogus statistics are a good reminder that people can be bright technicians, as well as ineligible education reformers.

I thought of Edison whenever I emphasized that educators need to adopt artificial intelligence as soon as possible to move away from the transformation of school and society.

Me, I study The history and future of education technologyAnd I have never seen an example of a school system – a country, state or municipality – which has quickly adopted a new digital technology and saw sustainable benefits for their students. The first districts of encouraging students to bring mobile phones to class did not make the youth better prepare better for the future than the more careful vision of schools. There is no evidence that their classrooms are the first countries to connect to the Internet, the economic growth, educational achievement or the good of the citizen.

New education technologies are just as strong as the community that guides their use. Easy to open a new browser tab; It is hard to make conditions for good education.

It takes years to develop new practices and criteria for academics, to take students new routines and to detect new supporting systems to improve education reliably for families. However, as the AI ​​schools are spread through the schools, both the Historical Tihasic analysis and new research operated with the K -12 teachers and students give some directions on navigating uncertainty and reducing damage.

We have become wrong and extra confident before

I started teaching high school history students for web searching in 20, at that time, library and information science experts developed an educational development for web evaluation that encouraged the websites to read the website for credibility markers: quote, appropriate format and a “about” page. We gave students such checklists Crump test – Currency, reliability, authority, accuracy and purpose – to guide their evaluation. We taught students to avoid Wikipedia and to believe with .com domains with .org or .edu domains and trust websites. All of them seemed reasonable and ignored.

The first Peer-Followed article displays the effective methods of teaching students How to search the web was published in 2019The It showed that the newcomers who usually use these taught techniques used to evaluate the skill of choosing the truth from the web fiction and perform badly in the tests. It has also shown that online information evaluation experts have used a completely different method: leaving a page quickly to see how other sources are featured. That method, now called FallQuickly, as a result of a more accurate search. The work was a bowl of bowel for an old teacher like me. We have spent on the ineffective ways of searching several million students for almost two decades.

Today, there is a cottage industry of consultants, original notors and “worried leaders” that travel to the country to train trainers on how to use AI in school. National and international organizations publish AI Literacy Frameworks that demanded that students need to know what skills are needed for the future. Technologists discover apps that encourage teachers and students to be used as a tutor, as a lesson planner, as an editor of writing or as part of conversation. As the method was invented today, the crump test proved as clear as it was.

There is better approach than to do extra confident estimates: to strictly examine new practices and techniques, and to support people who have only the evidence of effectiveness. Like web literacy, those evidences will take a decade or more to emerge.

But this time there is a difference. AI which I have said is a “Arrival technologyThe “AI is not invited to schools such as buying a desktop computer or smartboard – it crashes the party and that means schools will do something. 100 teachers across the United StatesAnd a great resentment is “Don’t make it alone to us.”

3 strategies for 3 cute paths

While waiting for a better answer from the education science community, which will take years, teachers have to be scientists themselves. I suggest three guideposts to proceed with AI in terms of uncertainty: humility, examination -audit and evaluation.

First, remind regular students and teachers that any school tries – literacy structure, education practice, new assessment – A best estimateThe In four years, students will hear that what they were first taught about the use of AI have been proved very wrong ever since. We all need to be ready to correct our thoughts.

Second, schools should check their students and curriculum and decide what kind of tests they want to do with AI. Some parts of your curriculum may invite sports and brave new efforts, others deserve more caution.

“Homework Machine,” in our podcast We have interviewed Eric TimansA teacher in California, a teacher who teaches a variant film making course. The final evaluation of his students is complex movies that require multiple technical and artistic skills to produce. An AI enthusiastic, Tims uses AI to develop its curriculum and he encourages students to use AI tools to solve filmmaking problems from scripting to technical design. He is not concerned about doing everything for AI students: as he says, “My students like to make movies. … then why would they replace AI?”

It is among the best, most thoughtful examples of the “everything” method I have encountered. I cannot even imagine a similar approach for a course like ninth grade English, where the main role of secondary school writing should probably be treated with a more careful approach.

Third, when teachers start a new test, they should be recognized that local evaluation will happen much faster than strict science. Every time schools introduce new AI policy or teaching practice, teachers should collect a pile of work related to the students who were created before using AI. If you allow students to use AI tools for constructive response in science labs, catch a pile of Surcha -2022 lab reports. Then, collect the new lab reports. Review whether post-i lab reports Show an improvement in the results you careAnd correct the practices accordingly.

Among local academics and international education scientists, people will learn a lot about AI in schools in schools by 2035. We could see that like the AI ​​Web, a place that is risky but in the end is important, full of useful resources that we continue to invite it to school. Or we can see that the AI ​​cellphones are like and negative effects Well -learning By the end the potential gains exceeds and thus are Make the best treatment with more aggressive restrictionsThe

Everyone in education feels urgent to solve the uncertainty around AI. But at first we don’t need any race to generate the answer – we need a competition to be correct.

JustinProfessor of Digital Media, Massachusetts Institute of Technology (MIT)

This article has been re -published from Conversation Under the Creative Commons License. Read KeyThe