Teaching Writing in the Age of AI
Acellus Writing Tutor Gold Edition arrives December 1. Here is how it works,why it works, and how it keeps every student’s writing their own.
Publish unlimited articles in the journal when you become a member. Ensure readers are seeing your latest insights and discoveries with the freedom to update your articles at any moment.
Instead of waiting a year or more for peer review, articles are published first and then open reviewed post-publication by members. All reviews, responses, and revisions are published along with the article.
Open access means no paywalls to read articles. No barrier to contributing to your field and getting cited.
Instead of a typical Article Processing Charge (APC) of $1,500-6,000+ to publish each article, members usually pay modest annual dues of $100 to publish unlimited articles.
Acellus Writing Tutor Gold Edition arrives December 1. Here is how it works,why it works, and how it keeps every student’s writing their own.
Generative artificial intelligence has made the finished essay an unreliable record of what a student knows. A student can now obtain a polished, on-topic composition in seconds, and an assessment that grades only the final product can no longer tell composition from transcription. This report describes Acellus Writing Tutor Gold Edition, a writing system for students in kindergarten through grade 12 that responds to this problem in two ways. First, it moves the object of instruction from the product to the process: students work through seven steps (Understand, Brainstorm, Draft, Revise, Edit, Publish, and Revise again), must correct every flagged spelling and grammar error themselves before a piece can be published, and then enter a scored revision loop in which a large language model (LLM) evaluates the writing against a rubric and the assignment’s requirements and coaches the student toward a stronger revision. Students may raise their score above 100 percent through bonus credit for continued revision. Second, it surrounds the writing with an authorship assurance architecture: work is permitted only inside the Acellus application, paste is disabled in the editor, devices are verified through platform attestation and per-install cryptographic keys, and a behavioral fingerprint built from session timing flags account sharing, outsourcing, and automation. We review the research basis for each design choice. The evidence is strong for explicit instruction in planning, revising, and editing; for formative feedback that names a next step; and for learning conditions in which students generate and retrieve answers themselves. Evidence for LLM-generated feedback on student writing is newer but encouraging, with the important qualification that unrestricted AI assistance can improve performance while impairing learning. Evidence for behavioral detection of misconduct is real but probabilistic: at realistic rates of misconduct, even accurate detectors produce many false flags. Several elements, including bonus credit above 100 percent, have no direct research base and should be treated as hypotheses. We identify vulnerabilities in LLM-based scoring (self-preference, verbosity bias, and prompt injection), assess the monitoring design against privacy law that governs minors’ data, and make twelve recommendations, the most important of which is to capture writing-process evidence inside the editor so that authorship is demonstrated by how a text was built rather than inferred from who was holding the device. We close with an evaluation plan.
AbstractEvery school district in the United States is bound by the same expectation: students must be taught the standards of the grade in which they are enrolled. Yet a large share of students arrive in each grade missingfoundational knowledge that should have been mastered one, two, or even several years earlier. Because district policy rightly resists moving students backward, teachers have historically been left to "patch andhope," re-teaching fragments of earlier grades while attempting to cover the current curriculum. Vectored Instruction®, introduced by Acellus roughly a decade ago and now embedded automatically in everyAcellus deployment, resolves this dilemma. When a student fails to demonstrate mastery of a grade-level lesson, a technology called Deficiency Diagnostics analyzes the pattern of the student's responses toidentify the specific missing prerequisite. A second component, Customized Personal Instruction, then reaches back as many grade levels as necessary to rebuild that foundation before returning the studentto the current lesson, all while the student remains enrolled in the grade-level course and on track for credit and graduation.This paper describes how Vectored Instruction works, summarizes the outcomes districts are reporting, and, for the first time, situates the method within the educational and cognitive science literature thatexplains why it works: learning hierarchies and the cumulative structure of knowledge, mastery learning, cognitive load theory, diagnostic assessment through distractor analysis, formative feedback, and theintelligent tutoring literature. It then offers practical guidance for teachers, curriculum designers, and textbook authors who wish to apply the principles of Vectored Instruction, explains why the method isparticularly indispensable for students studying independently at home, and identifies research-based directions for the next generation of the technology. A closing section introduces the CyberSchoolInitiative, through which local districts can sponsor supported home-study programs built on Acellus and Vectored Instruction.
The aim of the Journal is to advance science by making research easier to access and publish and by avoiding institutional bias or politics.
The Journal includes all scientific fields, including natural sciences, formal sciences, social sciences, and applied sciences like engineering and medicine. It does not include the humanities.
The Journal is primarily for scientific researchers and practitioners, but it includes some features to make it more accessible to laypeople.
The Journal publishes original research (research articles, brief reports, registered reports, monographs, theses, and dissertations), reviews (including systematic reviews and meta-analyses), cases (case reports, case series, case studies), methods (method articles, software tool articles, study protocols), and data notes (including genome notes). To counter publication bias, we welcome replications and null findings.
Articles are published on a continuous basis.
Because articles are published first and then peer reviewed, the Journal does not have a traditional editorial board that decides which articles to publish.
A preprint server archives articles before they are published and does not include a review process. The Journal’s articles are published first and then reviewed.
To make research immediately available, articles are published first, with post-publication peer review. Review is a transparent process, with all reviews and responses publicly visible. This gives reviewers credit for their work and discourages misbehavior such as personal attacks or coercive citation.
Because the Journal does not have a traditional editorial board that decides which articles to publish, it is not indexed in Web of Science, which Clarivate uses to calculate Journal Impact Factor.
Among other problems, Journal Impact Factor reflects the mean of citations of all a journal’s articles and can be skewed by a few highly cited articles. It is therefore not an accurate measure of the impact of a journal’s typical article. For this and other reasons, DORA recommends that
Each article shows views, downloads, saves, and shares.
The Journal does not charge article processing charges (APCs) to publish. Dues-paying members of the Academy who are of an appropriate tier may publish unlimited articles with no additional fees. For more information on membership, please see the membership tier comparison.
International Academy of Science (IAS), the publisher of the Journal, is a nonprofit organization, so the Journal is not driven by profit. IAS is supported by membership dues and donations. For more about IAS, visit www.science.edu/ias.
All members besides affiliates may publish articles. Members without advanced degrees may publish articles, but their articles will be held to the same standard of peer review for the scientific community.
Articles and reviews are copyrighted by their authors and licensed under the Creative Commons Attribution (CC-BY) license, which allows readers to read, download, and reuse articles and reviews as long as they give appropriate credit.
Using ANSI/NISO Standard Terminology for Peer Review, the Journal’s peer review is
All members may write reviews, but we ask reviewers to disclose relevant expertise and competing interests. A student or early-career researcher could have enough familiarity with the literature and field to write a peer review. For articles involving statistics, an expert statistician could write a statistical review, while a subject matter expert’s review could exclude statistics.
As part of a review, articles receive a star rating, which would invite abuse if a biased reviewer could anonymously leave 1-star or 5-star reviews that could not be viewed.
Reviewers should both stand behind their reviews and receive proper credit for their work.
If necessary to write the review, a reviewer may seek clarification from the article’s corresponding author in the comments section. For transparency, public comments are preferred over private messages and emails.
A review is a formal evaluation of an article’s rigor, writing, and contribution. It contributes to an article’s star rating. Authors often revise an article based on feedback from reviews.
A comment is informal and does not need to address all parts of the article. Comments can receive up or down votes based on their contribution to the scientific discussion.
While debate is welcome, all members are expected to follow the Code of Ethics.