Jocelyn Gecker

The No. 2 pencil became an American school staple through innovations such as yellow paint, attached erasers and standardized graphite grading. John Thoreau helped develop the grading system that made No. 2 pencils a balanced choice for writing and standardized tests, while the metal ferrule became another distinctive American feature. Although digital devices and online testing have reduced pencil use, handwritten exams are seeing renewed interest as schools seek to limit cheating and reduce screen exposure, preserving the pencil's cultural significance.

U.S. schools are moving from attempting to ban artificial intelligence to teaching students how to use it critically and safely. Charleston County, South Carolina, is training teachers and students to recognize AI hallucinations, bias, fabricated sources and privacy risks, while Utah has adopted a statewide strategy and trained thousands of educators. Supporters say AI literacy must include knowing when not to use these tools, but uneven funding and decentralized policies could widen disparities between districts.

U.S. schools are reassessing the rapid expansion of laptops, tablets and learning apps that accelerated during the COVID-19 pandemic. Los Angeles Unified has adopted one of the country’s most extensive restrictions, eliminating devices through second grade, imposing screen-time limits, blocking YouTube and banning devices during lunch and recess. Parent groups and educators argue that school-issued screens create distraction, encourage gaming and weaken reading, handwriting and in-person interaction, while districts also cite repair costs and inappropriate use. At least 14 states are considering screen-time legislation, although technology remains embedded in instruction for older students and some educators continue to view devices as necessary for access and learning.

College graduates at several US commencement ceremonies have booed speakers who emphasized artificial intelligence, reflecting widespread anxiety about AI-driven job losses and an already weak entry-level labor market. Eric Schmidt faced repeated jeers at the University of Arizona, while speakers at the University of Central Florida, Middle Tennessee State University and Marquette University also encountered student backlash. Polls indicate that about 70% of college students view AI as a threat to their job prospects, while Generation Z attitudes toward the technology are becoming more negative. Speakers argued that graduates can shape and use AI rather than fear it, but students criticized those messages as tone-deaf, particularly because many colleges restrict AI use while employers increasingly require AI-related skills.

A national analysis of test scores across more than 5,000 U.S. school districts finds that reading achievement has continued to decline since before the COVID-19 pandemic, leaving students about half a grade level behind pre-pandemic proficiency. Only five states and the District of Columbia recorded meaningful reading gains between 2022 and 2025, although math scores and attendance improved in most states. Louisiana, Maryland, Tennessee, Kentucky and Indiana were among the strongest reading performers, often after adopting phonics-based instruction and expanded support for struggling readers. Examples from Modesto and Detroit show that teacher training, targeted interventions and improved attendance can accelerate progress, but researchers stress that scores in many successful districts remain below grade level and that the causes of the broader decline remain contested.

College students are increasingly struggling to choose majors that will remain valuable as artificial intelligence changes education and employment. Many fear that entry-level work involving coding, analytics and other technical skills could be automated, while polling shows widespread concern about AI’s effect on job prospects. Some students are shifting toward marketing, liberal-arts subjects or other fields perceived to strengthen communication, critical thinking and interpersonal abilities, though healthcare and natural sciences may be less immediately affected. Educators and advisers acknowledge that they cannot confidently predict which skills will matter in the coming decades. Even computer science and data science students are anxious, despite some believing that AI expertise and the ability to communicate its complexities will create new opportunities.

Cornell German instructor Grit Matthias Phelps uses vintage manual typewriters for an “analog” assignment to prevent students from relying on generative AI, online translation tools and search engines. Students write without screens, spellcheckers or delete keys, forcing them to slow down, tolerate mistakes and seek help from classmates. The exercise also reduces digital distractions and encourages reflection on how technology has changed classroom interaction, though students face practical challenges such as slow typing and awkward corrections.

U.S. colleges are increasingly using oral exams, live discussions, presentations and other in-person assessments to determine whether students genuinely understand work increasingly produced with generative AI. Cornell, the University of Pennsylvania and New York University are adopting or studying oral formats, while NYU professor Panos Ipeirotis has developed an AI-powered oral exam that questions students about group projects. Educators say these methods can reinforce critical thinking, accountability and communication skills, though AI assessments can feel awkward and oral exams may challenge anxious students. Some students ultimately value the direct feedback and preparation for professional settings.

The Trump administration says an investigation by the Education Department’s Office for Civil Rights has prompted 31 universities to end partnerships with The PhD Project, a nonprofit that helps racial minorities pursue doctoral degrees. The department alleges that the organization unlawfully restricts eligibility by race and that participating institutions violated Title VI, while 14 additional universities remain in negotiations. Schools including MIT, the University of Michigan, Yale and others say they ended participation to avoid federal scrutiny; MIT resolved the matter without admitting wrongdoing. The department is also directing universities to review other partnerships for potentially race-based restrictions.

Brown University planned to release its delayed early-decision admission notices on Wednesday evening while authorities continued searching for the suspect in a campus shooting that killed two students and wounded nine. The university said it would work individually with admitted students who had concerns about enrolling. Admissions advisers said the shooting could worry some applicants but was unlikely to deter students strongly committed to Brown, where early-decision applications help fill a substantial share of the freshman class.

Colleges are increasingly using artificial intelligence to screen applications, read essays, review transcripts and summarize recommendations, although most say humans retain responsibility for admissions decisions. Virginia Tech is introducing an AI essay reader to help process tens of thousands of applications and accelerate decisions, while Georgia Tech and Stony Brook are using AI to reduce transcript-processing delays and errors. Caltech plans to use an AI video interview to assess whether applicants genuinely understand research they submit. The trend has prompted criticism, particularly at the University of North Carolina at Chapel Hill, and raised concerns about transparency, fairness, authenticity and student dignity. Admissions experts currently favor human oversight but warn that AI could eventually play a much larger role in deciding who is admitted.

Students should use AI tools such as ChatGPT to support learning rather than copy answers or outsource their thinking. Recommended uses include explaining difficult concepts, brainstorming, self-quizzing, receiving feedback and identifying knowledge gaps, while writing theses, drafts or essays for submission is generally discouraged. Students should check course-specific policies, discuss uncertain uses with instructors, disclose or cite AI assistance when appropriate, and follow academic-integrity and ethical standards. Policies vary widely among schools, making transparency and critical evaluation of AI-generated output essential.