AI won’t shortcut education. Unless we let it
A student opens a chapter. Types the question into AI. Copies the answer. Closes the book. Assignment: done.
This is the concern for educators across the world: the gradual outsourcing of thinking itself. The concern is valid – a 2024 Wharton study found students using general purpose AI to complete assignments showed weaker long-term outcomes. A separate analysis of millions of student interactions raised similar concerns about learners delegating higher-order thinking to AI.
Some students are also arriving at university without the reading skills they need. Of the 1.4 million students sitting the ACT, the United States’ primary college readiness assessment, only 39% met the Reading benchmark in 2025, down from 44% just four years earlier.
At the heart of this is active reading. Strategic highlighting, pausing to question, self-testing after a difficult chapter – these are strategies used by active readers to better comprehend what they read. They are the strongest predictors of genuine academic success. Yet, research suggests these skills are eroding.
Which raises a question: what if the same technology accelerating cognitive offloading could, under the right conditions, reverse it?
Over the past academic year, we analysed over 79 million student interactions with Pearson eTextbooks and instructor-led courseware. This was not general-purpose AI pulling answers from the open web. These tools were built on publisher-approved, expert-vetted content, designed around learning science principles – prompting students to question and test their own understanding rather than passively receive outputs.
What if the same technology accelerating cognitive offloading could, under the right conditions, reverse it
When AI is built this way, it does not replace active reading. It deepens it. In our research, a single interaction with a purpose-built AI study tool made a learner three times more likely to be classified as an active reader. Repeat use pushed that to 3.5 times. Inside instructor-led courseware, the effect was stronger still: 23 times more likely after a single session. Twenty-four times for repeat users.
But the finding that stayed with me was this: the students who gained most were those who had been the most passive readers to begin with. The ones every educator recognises – present but disengaged, going through the motions. Those students responded most to well-designed AI support.
For those of us working in international education, that direction matters. Students arriving at international universities from diverse linguistic backgrounds are already navigating a considerable gap between the texts they have encountered and the ones now in front of them. The tools we choose to give them are not neutral. A purpose-built study tool that allows students to converse with AI in over 100 languages, in their own words, about complex academic texts, begins to address that gap. One that doesn’t, compounds it.
For the student sitting in a lecture hall in a language that is not their first, navigating texts more complex than anything they have encountered before, the right tool at the right moment can be the difference between falling behind and finding their footing.
Read the latest Pearson AI in active learning research findings
About the author: Sharon Hague is a global growth strategist, passionate lifelong learner, and former educator with over 25 years of senior leadership experience at Pearson.
She currently serves as president of Pearson’s English language learning division, the company’s fastest-growing business unit, leading a global team. She is responsible for orchestrating the division’s strategy and driving a technology-powered growth plan, including market expansion, product diversification, and the development of AI-powered tools for corporations and educators.
Sharon also serves as Pearson’s UK CEO, acting as the company’s senior ambassador in its home market. Pearson is listed on the London Stock Exchange and is a constituent of the FTSE 100 Index.
Sharon’s previous roles include managing director of School Qualifications and School Assessment, where she led global sales and operations for Pearson’s K-12 businesses. She has worked extensively with governments, schools, and partners to deliver services that help learners make progress in their lives.
The post AI won’t shortcut education. Unless we let it appeared first on The PIE News.