Liza Libes, writing in Minding the Campus in April 2026 and in her Substack publication Pens & Poison, made a specific claim about the graduating class of 2026: they are the first cohort to have passed through all four years of high school with AI writing tools available throughout. The implication is structural, not anecdotal. These students have never had to write an entire essay from scratch. They have never had to map out a cohesive argument starting from nothing. And the observable result, in college classrooms that have begun receiving them, is not what many expected. It is not that students submit AI-generated essays. It is that students have internalized the AI cadence as their default writing register. Asked to write unassisted, they produce AI-sounding sentences without the tool.
Libes identifies a specific tell: an overwhelming number of student essays feature the construction "It's not just X, it's Y" — one of ChatGPT's signature formulations. Students produce this without AI open in front of them. They have learned to write the way AI writes because that is the model they have been reading and, in many cases, revising. The result is a student who can produce a grammatically sound paragraph that reads like something worth saying, without having done the work of actually forming an idea. The surface looks like writing. The thinking underneath it is absent.
For ELA teachers, this is not an argument for banning AI. It is an argument for redesigning the conditions under which students write. A student who never writes under constraint, without AI, without notes, under time pressure, never develops the capacity to generate original thought on demand. That capacity is what writing instruction is supposed to produce. If assignments can be completed with AI, they will be, and the student who completes them that way is practicing text production without practicing thinking. Those are not the same skill.
Ask students to write one original claim about today's content in 3 minutes, by hand, no starters, no template. Read four or five back to the class anonymously. Then ask: which of these is something the writer actually thinks, versus something that sounds like what you're supposed to say? The distinction most students can hear it when it's read aloud. Making them hear it is the first move toward teaching them to produce the real version. Any subject, any grade level.
The College Board published research in February 2026 based on a survey of more than 3,000 college faculty. The headline finding: 84% of faculty agree that AI tools make students more dependent on technology and less likely to develop critical thinking skills, express original ideas, or engage deeply with course material. A secondary finding is nearly as significant: 79% of faculty describe themselves as either "just starting to explore" what guidance on AI use looks like or as having "some understanding but still needing further guidance." Faculty teaching writing-intensive courses report the highest rates of challenge. Faculty who have themselves experimented with AI tools hold more favorable views of its potential than faculty who have not. The College Board research was conducted before the fall 2026 semester brought the first AI-native graduating cohort — students who completed all four years of high school with these tools — into college classrooms.
The 84% figure is not a faculty complaint about lazy students. It is a description of a skill substitution that has already happened. When a student uses AI to generate the first draft of every essay, they are not building the skill of idea generation. That skill is what essay writing is supposed to develop. The College Board data does not measure your current students directly, but the class of 2026 arriving in colleges this fall is the same cohort your current juniors and seniors belong to. If the pattern holds, your students who have been using AI throughout high school are carrying the same substitution into your classroom. The implication for ELA specifically is precise: if writing assignments can be completed with AI, and students complete them that way, the assignment is not developing what it was designed to develop. That is an instructional problem, not a technology problem.
The 74 Million reported in January 2026 on research examining Science of Reading implementation across four major urban school districts. The finding is specific and uncomfortable: students in these districts demonstrated stronger foundational skills — phonics, decoding, basic fluency — but the research found evidence that the approach "may unintentionally encourage teachers to focus on surface-level goals" rather than deeper literacy. The researchers used the phrase "robust comprehension" to describe what was missing: the ability to interpret, infer, and construct sustained meaning from complex text. A separate 2026 study from SRI International, covering large urban districts in Texas, Maryland, Virginia, and North Carolina, found that between 72% and 89% of teachers in high-implementation districts were using comprehension-focused materials daily. The two findings together describe a field that has made real progress on one problem and is now encountering the next one.
The Science of Reading has largely won the policy argument. More than 40 states have introduced SoR-aligned legislation in the last five years. What The 74's January reporting describes is the limit of that victory: winning on decoding does not automatically produce the deeper reading skills that high school teachers need students to have. Students who arrive in 9th grade as fluent decoders but weak comprehenders have always existed. The SoR movement is now producing more of them, at scale, in districts that implemented the approach correctly. For high school ELA and content-area teachers, that is the incoming population. The diagnostic question is not "Can this student read the words?" but "Can this student say what the text is doing, why the author made those choices, and what a counterargument would require?"
The Massachusetts House of Representatives passed a bill 129-25 on April 9, 2026 that combines two policy instruments that had been traveling separately in other states: a bell-to-bell ban on cellphones in public schools and a prohibition on social media access for children under 14. The bill also requires parental consent for 14- and 15-year-olds to create social media accounts. The measure, which had passed the Senate in a narrower form covering only the phone ban, now returns to the Senate for reconciliation. If passed and signed by Governor Maura Healey, it takes effect October 1, 2026. LGBTQ+ advocates have raised specific concerns about the parental data-access provisions for 14- and 15-year-olds, arguing they could force disclosure of sensitive information to unsupportive families. The American Civil Liberties Union of Massachusetts has expressed concern that platforms may block access to legitimate resources, including health and support information, under a broad social media definition.
The Massachusetts bill is the most consequential state-level youth digital legislation passed in 2026. It is significant not because of the phone ban — which is by now familiar territory — but because it combines the phone ban with an age-based social media prohibition in the same legislation. That combination frames school phone policy not just as an attention or academic performance problem, but as part of a larger claim that youth access to social media at current levels constitutes a social harm serious enough to warrant legal restriction outside the school building. For teachers in Massachusetts, an October 1 effective date means implementation planning during back-to-school season. For teachers everywhere else, the bill provides the legislative template that other states will cite in 2027 sessions.