New digital-reading evidence is sharpening an old warning: screens are not automatically bad for learning, but the way text is displayed and the way students move through it can make deep comprehension harder. The practical issue is not whether a learner reads on paper or on a device. It is whether the reading setup slows the learner down enough to build a mental model, question the text, and retrieve the main ideas without looking.
The clearest recent data point is a 2026 network meta-analysis in Education and Information Technologies, Decoding digital reading: a network meta-analysis of comprehension across devices. The researchers synthesized 56 studies and compared paper with computers, tablets, e-readers, and smartphones. Across all comparisons, comprehension was stronger on paper than on computers and smartphones, while tablets and e-readers were closer to paper. The authors also highlighted scrolling as an important design issue: when text keeps moving, readers may lose the spatial cues that help them remember where ideas appeared and how the argument was organized.
This matters because much school and college reading now happens in formats that encourage speed: scrolling pages, notifications nearby, easy tab switching, searchable summaries, and short excerpts. A learner may finish the page and still have a weak grasp of the structure. Digital reading can feel efficient while quietly reducing the amount of active monitoring the reader does.
The same concern shows up in newer research on AI-supported academic reading. In the 2026 preprint Self-Regulated Reading with AI Support: An Eight-Week Study with Students, researchers followed 15 undergraduates across 239 reading sessions and analyzed 838 prompts. Students often used AI for comprehension support, but metacognitive prompts were much less common. The researchers described a pattern in which students sometimes read through AI-generated summaries rather than with the original text, using summaries to decide which sections deserved closer attention.
That pattern is important for deep comprehension. Summaries can help a reader preview structure, check a confusing passage, or compare interpretations. But if the summary becomes the primary reading material, the learner may skip the hard work that builds durable understanding: identifying claims, tracking evidence, noticing confusion, and explaining the argument in their own words.
Education news is also moving in this direction. The Hechinger Report recently covered parent and educator pushback against routine screen use in early grades in Parents are pushing back on too much screen time for kids in school. A separate Hechinger commentary argued that the answer is not to abandon technology, but to hold it to a higher standard: use digital tools when they add something print cannot, such as immediate feedback, simulation, visualization, or authentic interaction: Walking away from education technology is not the answer.
Assessment policy is catching up too. OECD’s 2026 PISA announcement says the forthcoming PISA 2025 results will include new evidence on students’ AI use for schoolwork and a new Learning in the Digital World domain that examines computational problem solving and self-regulated learning: New PISA results coming soon. That is a signal that digital reading is now part of a wider learning question: can students manage attention, strategy, effort, and judgment when the learning environment is online?
What Happened
Recent research and education reporting are converging on a more precise view of digital reading. Paper still has advantages for many comprehension tasks, especially longer informational reading. But device type, text layout, scrolling, annotation tools, and AI assistance all change the learning outcome. The problem is not merely screen exposure. The problem is shallow processing.
In practice, a student reading a long article on a phone, scrolling continuously, and checking an AI summary may be doing a very different cognitive task from a student reading the same article on a tablet with stable pages, margin notes, pauses, and self-testing. Both are “digital reading,” but only one is likely to support deep comprehension.
Why It Matters
Deep comprehension depends on more than recognizing sentences. Learners need to connect ideas, identify the author’s purpose, distinguish evidence from examples, and remember the main structure after the text is gone. Digital environments can weaken those habits when they reward completion, search, skimming, and quick answers.
The risk is overconfidence. A learner can feel fluent because the text is searchable, highlighted, summarized, or still open on the screen. But real comprehension shows up when the learner can close the page and explain the argument accurately, answer questions from memory, and apply the idea to a new case.
The Practical Learning Conclusion
Use digital reading deliberately. When the goal is deep comprehension, design the session so the screen behaves more like a thinking space and less like a feed.
- Slow the text down. For long or difficult readings, use a larger screen, stable page view, or print when possible. Avoid reading complex material in a narrow phone layout unless there is no alternative.
- Pause after each section. Before scrolling on, write one sentence that states the section’s main claim. If the claim is vague, reread before continuing.
- Question the argument. Add margin notes or comments that ask: What is the author claiming? What evidence supports it? What would count as a counterexample?
- Use AI as a check, not a shortcut. First explain the passage yourself. Then use a summary or chatbot to compare, clarify, or generate questions. Do not let the summary replace the original reading.
- Retrieve the main ideas after reading. Close the tab and write three bullets from memory: the main point, the strongest evidence, and one implication. Reopen the text only after attempting recall.
- Turn highlights into tasks. Highlighting is not comprehension by itself. Convert each important highlight into a question you can answer later without the text.
- Check for transfer. After reading, apply the idea to a new example, problem, policy, experiment, or personal study routine. Transfer reveals whether the idea was understood or merely recognized.
The learning conclusion is simple: digital reading needs friction. Learners should build pauses, questions, annotations, and retrieval into the session so the device does not pull them into passive scrolling. Whether the text is on paper, tablet, laptop, phone, or inside an AI-supported workflow, deep comprehension still requires slowing down, asking better questions, and proving the main ideas from memory.