Tech giants racing to embed AI in schools around the globe

In early November, Microsoft said it would supply artificial intelligence tools and training to more than 200,000 students and educators in the United Arab Emirates.

Days later, a financial services company in Kazakhstan announced an agreement with OpenAI to provide ChatGPT Edu, a service for schools
and universities, for 165,000 educators in Kazakhstan.

Last month, xAI, Elon Musk’s AI company, announced an even bigger project with El Salvador: developing an AI tutoring system, using the company’s Grok chatbot, for more than one million students in thousands of schools there.

Fuelled partly by American tech companies, governments around the globe are racing to deploy generative AI systems and training in schools and universities.

Some US tech leaders say AI chatbots — which can generate humanlike emails, create class quizzes, analyse data and produce computer code — can be a boon for learning. The tools, they argue, can save teachers time, customise student learning and help prepare young people for an “AI-driven” economy.

But the rapid spread of the new AI products could also pose risks to young people’s development and well-being, some children’s and health groups warn.

A recent study from Microsoft and Carnegie Mellon University found that popular AI chatbots may diminish critical thinking. AI bots can produce authoritative-sounding errors and misinformation, and some teachers are grappling with widespread AI-assisted student cheating.

Silicon Valley for years has pushed tech tools like laptops and learning apps into classrooms, with promises of improving education access and revolutionising learning.

Still, a global effort to expand school computer access — a programme known as “One Laptop per Child” — did not improve students’ cognitive skills or academic outcomes, according to studies by professors and economists of hundreds of schools in Peru. Now, as some tech boosters make similar education access and fairness arguments for AI, children’s agencies like UNICEF are urging caution and calling for more guidance for schools.

“With One Laptop per Child, the fallouts included wasted expenditure and poor learning outcomes,” Steven Vosloo, a digital policy specialist at UNICEF, wrote in a recent post. “Unguided use of AI
systems may actively de-skill students
and teachers.”

Education systems across the globe are increasingly working with tech companies on AI tools and training programs.

In the United States, where states and school districts typically decide what to teach, some prominent school systems recently introduced popular chatbots for teaching and learning. In Florida alone, Miami-Dade County Public Schools, the nation’s third-largest school system, rolled out Google’s Gemini chatbot for more than 100,000 high school students. And Broward County Public Schools, the nation’s sixth-biggest school district, introduced Microsoft’s Copilot chatbot for thousands of teachers and staff members.

Outside the United States, Microsoft in June announced a partnership with the Ministry of Education in Thailand to provide free online AI skills lessons for hundreds of thousands of students. Several months later, Microsoft said it would also provide AI training for 150,000 teachers in Thailand. OpenAI has pledged to make ChatGPT available to teachers in government schools across India.

The Baltic nation of Estonia is trying a different approach, with a broad
new national AI education initiative called “AI Leap.”

The programme was prompted partly by a recent poll showing that more than 90% of the nation’s high-schoolers were already using popular chatbots like ChatGPT for schoolwork, leading to worries that some students were beginning to delegate school assignments to AI.

Estonia then pressed US tech giants to adapt their AI to local educational needs and priorities. Researchers at the University of Tartu worked with OpenAI to modify the company’s Estonian-language service for schools so it would respond to students’ queries with questions rather than produce direct answers.

Introduced this school year, the “AI Leap” programme aims to teach educators and students about the uses, limits, biases and risks of AI tools. In its pilot phase, teachers in Estonia received training
on OpenAI’s ChatGPT and Google’s Gemini chatbots.

“It’s critical AI literacy,” said Ivo Visak, the chief executive of the AI Leap Foundation, an Estonian nonprofit that is helping to manage the national education programme. “It’s having a very clear
understanding that these tools can be
useful — but at the same time these tools can do a lot of harm.”

Estonia also recently held a national training day for students in some high schools. Some of those students are now using the bots for tasks like generating questions to help them prepare for school tests, Visak said.

“If these companies would put their effort not only in pushing AI products, but also doing the products together with the educational systems of the world, then some of these products could be really useful,” Visak added.

This school year, Iceland started its own national AI pilot in schools. Now several hundred teachers across the country are experimenting with Google’s Gemini chatbot or Anthropic’s Claude for tasks like lesson planning, as they aim to find helpful uses and to pinpoint drawbacks.

Researchers at the University of Iceland will then study how educators used the chatbots.

Students won’t use the chatbots for now, partly out of concern that relying on classroom bots could diminish important elements of teaching and learning.

“If you are using less of your brain power or critical thinking — or whatever makes us more human — it is definitely not what we want,” said Thordis Sigurdardottir, the director of Iceland’s Directorate of Education and School Services.

Tinna Arnardottir and Frida Gylfadottir, two teachers participating in the pilot at a high school outside Reykjavik, say the AI tools have helped them create engaging lessons more quickly.

Arnardottir, a business and entrepreneurship teacher, recently used Claude
to make a career exploration game to
help her students figure out whether they were more suited to jobs in sales, marketing or management. Gylfadottir, who teaches English, said she had uploaded some vocabulary lists and then used
the chatbot to help create exercises for
her students.

“I have fill-in-the-blank word games, matching word games and speed challenge games,” Gylfadottir said. “So before they take the exam, I feel like they are better prepared.”

Gylfadottir added that she was concerned about chatbots producing misinformation, so she vetted the AI-created games and lessons for accuracy before asking her students to try them. Gylfadottir and Arnardottir said they also worried that some students might already be growing dependent on — or overly trusting of — AI tools outside school.

That has made the Icelandic teachers all the more determined, they said, to help students learn to critically assess and use chatbots.

“They are trusting AI blindly,” Arnardottir said. “They are maybe losing motivation to do the hard work of learning, but we have to teach them how to learn with AI.”

Teachers currently have few rigorous studies to guide generative AI use in schools. Researchers are just beginning to follow the long-term effects of AI chatbots on teenagers and schoolchildren.

“Lots of institutions are trying AI,” said Drew Bent, the education lead at Anthropic. “We’re at a point now where we need to make sure that these things are backed by outcomes and figure out what’s working and what’s not working.”