What might teach us the previous educational technology failures about the future of artificial intelligence in schools

This article was originally published on Conversation.

American technicians have told teachers to adopt their new inventions for more than a century. In 1922, Thomas Edison announced In the near future, all the textbooks will be replaced by films, because the text was 2 % effective, but the film was 100 % effective. These false statistics are a good reminder that people can be great technicians, while educational reformists are incompetent.

I think of Edison whenever I hear technicians who insist that teachers must adopt artificial intelligence as soon as they can advance the shift on the verge of washing schools and society.

At the Massachusetts Institute of Technology, I study History and future of educational technologyAnd I never faced an example of a school system – a country, country or municipality – quickly adopted a new digital technology and witnessed strong benefits for their students. The first regions that encourage students to bring mobile phones to class have not been better than preparing young people for the future than schools that followed a more cautious approach. There is no evidence that the first countries to connect their classroom to the Internet stand in economic growth, educational attainment, or citizen’s welfare.

New education technologies are only powerful as societies that direct their use. It is easy to open the new browser tab; It is difficult to create conditions for good learning.

It takes years for teachers to develop new practices and rules, for students to adopt new measures, and for families to determine new support mechanisms for a new invention to improve learning reliably. But with the spread of artificial intelligence through schools, both the historical analysis and the new research that was conducted with K-12 teachers and students provide some directives about moving in uncertainty and reducing damage.

We were wrong and made a mistake before

She started teaching high school history students to research the web in 2003. At that time, experts in library and information science science sciences have developed the teaching assets to evaluate web that encouraged students to read web sites closely in search of signs of credibility: the appropriate categories and coordination and page “about”. We gave students examination lists like CRAP test Currency, reliability, power, accuracy and purpose – to direct their evaluation. Students taught us to avoid Wikipedia and trust in web sites using the .org or .EDU fields in the fields of .com. Everything seemed reasonable and enlightened at that time.

The first article reviewed by the peers showing effective methods of teaching students How to search the web was published in 2019. It has been shown that beginners who used these techniques that are commonly taught are miserably failed in the tests that evaluate their ability to sort the truth from the imagination on the web. It also showed that experts in online information evaluated a completely different approach: left a page quickly to learn how to describe other sources. This method is now called Side readingIt resulted in faster and more accurate research. The work was an intestinal punch for an old teacher like me. We have spent nearly two decades teaching millions of students ineffective ways of research.

Today, there is a cottage industry of consultants, the main matches and “thought leaders” who are traveling to the country claiming to train teachers on how to use artificial intelligence in schools. National and international organizations publish reading and writing frameworks, Amnesty International, which claims to know the skills that students need for their future. Technicians invent applications that encourage teachers and students to use childbirth intelligence as teachers, such as lessons, such as writing editors, or as conversation partners. These methods have a lot of support for proof today as did the CRAP test when it was invented.

There is a better approach than excessive confidence: a strictly testing new practices and strategies and a large -scale call for people who have strong evidence of effectiveness. As with literacy on the Internet, this evidence will take one or more decades to appear.

But there is a difference this time. Artificial intelligence is what I called “Access technology“Artificial intelligence is not invited to schools through the adoption process, such as buying a desktop computer – which disrupts parties, however, schools must do something. 100 teachers from all over the United StatesAnd widespread abstinence is “Do not make us go alone.”

3 strategies for the wise path forward

While waiting for better answers than the Education Science Society, which will take years, teachers must be scientists themselves. I recommend three guidelines to move forward with artificial intelligence under uncertainty conditions: humility, experimentation and evaluation.

First, students and teachers regularly remind that anything that schools are trying – reading and writing frameworks, teaching practices, and new assessments – It is the best guess. Within four years, students may hear that what was taught for the first time about the use of artificial intelligence has proven to be completely wrong. We all need to be ready to review our thinking.

Second, schools need to examine their students and curricula, and determine the types of experiences they want to conduct with artificial intelligence. Some parts of your curriculum may invite fun and bold new efforts, while others deserve more caution.

In our podcast, “Home Duty Machine”, We met Eric TimonzA teacher in Santa Anna, California, who teaches optional filmmaking courses. The final assessments of his students are complex films that require multiple technical and technical skills to produce them. Timmons, one of the artificial intelligence lovers, uses Amnesty International to develop its curriculum, and encourages students to use artificial intelligence tools to solve film industry problems, from text programming to artistic design. It is not worried about doing everything for students: He says, “My students love to make films … so why does this be replaced by Amnesty International?”

It is among the best and most studied examples of the “All In” approach. I can also not imagine the recommendation of a similar approach to a ninth -class course, as the central introduction to secondary school writing should be dealt with with a more careful approach.

Third, when teachers launch new experiences, they must realize that local evaluation will happen much faster than strict science. Each time schools launch a new policy or teaching Amnesty International, teachers must collect a pile of students related to being developed before the use of artificial intelligence during teaching. If you allow students to use artificial intelligence tools for feedback on science laboratories, get a pile of Circa-2012 laboratory reports. Then, collect the new laboratory reports. See whether AI’s laboratory reports Show an improvement in the results you care aboutAnd review practices accordingly.

Among the local teachers and the international community of education scientists, people will learn a lot by 2035 about artificial intelligence in schools. We may find that artificial intelligence is similar to web, which is a place with some risks, but in the end it is full of important and useful resources that we continue to invite to school. Or we may find that artificial intelligence resembles mobile phones, and the negative effects on Welfare and learning In the end, the possible gains exceeded, and therefore it is The best processing with more aggressive restrictions.

Everyone in education is urgent to resolve uncertainty about artificial intelligence. But we do not need a race to create answers first – we need a race to be right.

Justin RishProfessor of digital media, Massachusetts Institute of Technology (MIT)

This article has been republished from Conversation Under the creative public license. Read The original article.