essay feedback ai

• Skip to main content
• Skip to primary sidebar
• Skip to footer

Additional menu

Khan Academy Blog

Introducing Khanmigo’s New Academic Essay Feedback Tool

posted on November 29, 2023

By Sarah Robertson , senior product manager at Khan Academy

Khan Academy has always been about leveraging technology to deliver world-class educational experiences to students everywhere. We think the newest AI-powered feature in our Khanmigo pilot—our Academic Essay Feedback tool—is a groundbreaking step toward revolutionizing how students improve their writing skills.

The reality of writing instruction

Here’s a word problem for you: A ninth-grade English teacher assigns a two-page essay to 100 students. If she limits herself to spending 10 minutes per essay providing personalized, detailed feedback on each draft, how many hours will it take her to finish reviewing all 100 essays?

The answer is that it would take her nearly 17 hours —and that’s just for the first draft!

Research tells us that the most effective methods of improving student writing skills require feedback to be focused, actionable, aligned to clear objectives, and delivered often and in a timely manner . 

The unfortunate reality is that teachers are unable to provide this level of feedback to students as often as students need it—and they need it now more than ever. Only 25% of eighth and twelfth graders are proficient in writing, according to the most recent NAEP scores .

An AI writing tutor for every student

Developed by experts in English Language Arts (ELA) and writing instruction, the pilot Khanmigo Academic Essay Feedback tool uses AI to offer students specific, immediate, and actionable feedback on their argumentative, expository, or literary analysis essays. 

Unlike other AI-powered writing tools, the Academic Essay Feedback tool isn’t limited to giving feedback on sentence- or language-level issues alone, like grammar or spelling. Instead, it provides feedback on areas like essay structure and organization, how well students support their arguments, introduction and conclusion, and style and tone.

The tool also doesn’t just stop at providing feedback, it also guides students through the revision process. Students can view highlighted feedback, ask clarifying questions, see exemplar writing, make revisions, and ask for further review—without the AI doing any actual writing for them.

Unique features of Khanmigo pilot Academic Essay Feedback tool

• Immediate, personalized feedback: within seconds, students get detailed, actionable, grade-level-appropriate feedback (both praise and constructive) that is personalized to their specific writing assignment and tied directly to interactive highlights in their essay.
• Comprehensive approach: feedback covers a wide range of writing skills, from crafting an engaging yet focused introduction and thesis, to overall essay structure and organization, to style and tone, to alignment and use of evidence.
• Interactive revision process: students can interact with Khanmigo to ask questions about specific pieces of feedback, get examples of model writing, make immediate revisions based on the feedback, and see if their revisions addressed the suggestion.
• Support for various essay types: the tool is versatile and assists with multi-paragraph persuasive, argumentative, explanatory, and literary analysis essay assignments for grades 8-12 (and more, coming soon).
• Focus on instruction and growth: like all Khanmigo features, the Academic Essay Feedback tool will not do the work for the student. Teachers and parents can rest assured that Khanmigo is there to improve the students’ independent writing skills, not provide one-click suggested revisions.

How parents can use Khanmigo’s Academic Essay Feedback tool

Any student with Khanmigo access can find the feedback tool under the “Write” category on their AI Activities menu. 

For academic essays, students should simply paste their first draft into the essay field, select their grade level and essay type, and provide the essay instructions from the teacher.

Students then click “Submit” and feedback begins generating. Once Khanmigo is done generating feedback, students can work their way through the suggestions for each category, chat with Khanmigo for help, make revisions, and resolve feedback. They can then submit their second draft for another round of feedback, or copy the final draft to submit to their teacher.

Bringing Khanmigo to your classroom, school, or district

Teachers in Khan Academy Districts partnerships can begin using the Khanmigo Academic Essay Feedback tool with their students right away. Simply direct students to the feedback tool under the “Write” category on their AI Activities menu.

Like all other Khanmigo activities, students’ interactions are monitored and moderated for safety. Teachers or parents can view the student’s initial draft, AI-generated feedback, chat history, and final draft in the student’s chat history. If anything is flagged for moderation, teachers or parents will receive an email notification.

Looking ahead

With the Academic Essay Feedback tool in our Khanmigo pilot, teachers and parents can empower students to take charge of their writing.The tool helps facilitate a deeper understanding of effective writing techniques and encourages self-improvement. For teachers, we think this tool is a valuable ally, enabling them to provide more frequent, timely, detailed, and actionable feedback for students on multiple drafts.

In the coming months, we’ll be launching exciting improvements to the tool and even more writing resources for learners, parents, teachers, and administrators:

• The ability for teachers to create an essay-revision assignment for their students on Khan Academy
• More varied feedback areas and flexibility in what feedback is given
• Support for students in essay outlining and drafting
• Insights for teachers and parents into their students’ full writing process

Stay tuned!

Sarah Robertson is a senior product manager at Khan Academy. She has a M.Ed. in Curriculum and Instruction and over a decade of experience teaching English, developing curriculum, and creating software products that have helped tens of millions of students improve their reading and writing skills.

Get Khanmigo

The best way to learn and teach with AI is here. Ace the school year with our AI-powered guide, Khanmigo. 

For learners     For teachers     For parents

Automated Feedback

Elevate and optimize the feedback on academic writing.

Automated Feedback addresses mechanical and structural elements like citation, academic style, grammar, and structure, leaving teachers with more time to focus on higher-order cognitive skills like comprehension and critical argumentation of concepts.   ‍ Read more about the iterative feedback system ↗ Automated Feedback tool helps build a holistic learning experience over a correctional one by taking care of lower-order concerns, making personalized feedback possible even for a large student cohorts .

Automated Feedback saves time with instant feedback

CREATE SCALABLE AI-POWERED FEEDBACK OPPORTUNITIES

Give more students more feedback in less time, making personalized feedback possible at even the largest scales. 

STREAMLINE LOWER-ORDER FEEDBACK

Provide students with real-time, formative feedback on their academic writing skills based on criteria selected by instructors.

OVERVIEW of learning analytics

Access class-wide and individual learning analytics via built-in dashboard that highlight larger points of improvement, enabling opportunities for personalized learning experiences.

Tool features

Customize your selection of criteria based on classroom contexts and specific writing assignments. Once selected, students will receive feedback on these criteria in real-time. You can choose criteria related to content, language, layout, references, and figures and tables.

Upload documents and Try-Out your selected criteria before providing it to students to ensure the feedback is in line with their learning goals. 

Choose from suggested templates (essay, thesis, research paper) catered towards specific writing assignments, and further customize the recommended set of criteria, adjusting it to suit your needs.

Instant class-wide analytics now leaves you with the opportunity to guide students in specific areas of improvement. View to what extent students met each criterion, and assess individual students to help clear any doubts.

Enhance student’s learning process by adding Automated Feedback as an additional module to our Peer Review or Assignment Review tools, creating additional sources of feedback. 

Choose from suggested templates (essay, case study, research paper) catered towards specific writing assignments, and further customize the recommended set of criteria, adjusting it to suit your teaching and learning goals.

Instant class-wide analytics gives you the opportunity to guide students in improving specific aspects of their assignments and writing skills. View to what extent each student met the criteria, and reach out to support them.

Enhance students' learning process by adding Automated Feedback as an additional module to our Peer Review or Assignment Review tools, adding a another layer of feedback.

If the institution has a license with Turnitin, instructors can enable plagiarism check setting for students to have their submissions reviewed for plagiarism by Turnitin before the automated feedback activity.

Group Formation works in your LMS, so you and your students don't need a separate login account and do the activity with ease. See integrations →

Your data is hosted by Microsoft Azure in your region. We're compliant with data protection laws including the GDPR and WCAG 2.1 (AA) accessibility standards.

Our team offers teachers and students 24/5 human support, help articles and workshops to familiarize instructors with FeedbackFruits.

Assignment Review

Deliver quality feedback on students' assignments in less time with Assignment Review.

Peer Review

Streamline effective peer feedback process with Peer Review.

Instructors love FeedbackFruits

“Ultimately I'd like to provide detailed feedback for every single assignment but that's unrealistic. Automated Feedback did something I couldn't provide for students.”

"I really enjoyed [Automated Feedback] because it does exactly what I hoped it would - it takes out the minor issues and lets the peers look at the text at a higher level."

"My enthusiasm over it is that it’s fantastic! We will definitely keep using Automated Feedback in our online classes."

"[Automated Feedback is] a very useful tool to take away burden from lecturers and supervisors and increase the feedback that is provided to students."

"I noticed that learning of students improved. First, they were enthusiastic about the tool and also interaction between students really helped them to better prepared for the subject."

The fruit of Edtech Dotank co-creation

Explore the FeedbackFruits all-in-one Teaching & Learning System

AI Writing Reviewer

Ai-powered writing review and feedback tool.

• Improve your blog posts: Get feedback on your drafts to improve clarity, structure, and engagement.
• Elevate your academic writing: Receive suggestions to enhance your thesis, research papers, or essays.
• Polish your professional emails: Make sure your emails are clear, concise, and effective.
• Refine your creative writing: Get insights on how to improve the flow, character development, and plot structure in your stories.
• Enhance your marketing content: Improve the persuasiveness and impact of your marketing messages.

New & Trending Tools

Ai speech generator, detailed lesson plan builder, love letter generator.

EssayGrader

The world’s leading AI platform for teachers to grade essays

EssayGrader is an AI powered grading assistant that gives high quality, specific and accurate writing feedback for essays. On average it takes a teacher 10 minutes to grade a single essay, with EssayGrader that time is cut down to 30 seconds. That's a 95% reduction in the time it takes to grade an essay, with the same results.

How we've done

Happy users

Essays graded

EssayGrader analyzes essays with the power of AI. Our software is trained on massive amounts of diverse text data, including books, articles and websites. This gives us the ability to provide accurate and detailed writing feedback to students and save teachers loads of time. We are the perfect AI powered grading assistant.

EssayGrader analyzes essays for grammar, punctuation, spelling, coherence, clarity and writing style errors. We provide detailed reports of the errors found and suggestions on how to fix those errors. Our error reports help speed up grading times by quickly highlighting mistakes made in the essay.

Bulk uploading

Uploading a single essay at a time, then waiting for it to complete is a pain. Bulk uploading allows you to upload an entire class worth of essays at a single time. You can work on other important tasks, come back in a few minutes to see all the essays perfectly graded.

Custom rubrics

We don't assume how you want to grade your essays. Instead, we provide you with the ability to create the same rubrics you already use. Those rubrics are then used to grade essays with the same grading criteria you are already accustomed to.

Sometimes you don't want to read a 5000 word essay and you'd just like a quick summary. Or maybe you're a student that needs to provide a summary of your essay to your teacher. We can help with our summarizer feature. We can provide a concise summary including the most important information and unique phrases.

AI detector

Our AI detector feature allows teachers to identify if an essay was written by AI or if only parts of it were written by AI. AI is becoming very popular and teachers need to be able to detect if essays are being written by students or AI.

Create classes to neatly organize your students essays. This is an essential feature when you have multiple classes and need to be able to track down students essays quickly.

Our mission

At EssayGrader, our mission is crystal clear: we're transforming the grading experience for teachers and students alike. Picture a space where teachers can efficiently and accurately grade essays, lightening their workload, while empowering students to enhance their writing skills. Our software is a dynamic work in progress, a testament to our commitment to constant improvement. We're dedicated to refining and enhancing our platform continually. With each update, we strive to simplify the lives of both educators and learners, making the process of grading and writing essays smoother and more efficient.We recognize the immense challenges teachers face – the heavy burdens, the long hours, and the often underappreciated efforts. EssayGrader is our way of shouldering some of that load. We are here to support you, to make your tasks more manageable, and to give you the tools you need to excel in your teaching journey.

Join the newsletter

Subscribe to get our latest content by email.

Super  charge your college essay

Elevate Your Essay to Perfection

Admissions expertise

Impression analysis

Actionable insights

Instant feedback

Get started free.

Recent blog posts

Frequently Asked Questions

Revolutionize Your Writing Process with Smodin AI Grader: A Smarter Way to get feedback and achieve academic excellence!

For Students

Stay ahead of the curve, with objective feedback and tools to improve your writing.

Your Virtual Tutor

Harness the expertise of a real-time virtual teacher who will guide every paragraph in your writing process, ensuring you produce an A+ masterpiece in a fraction of the time.

Unbiased Evaluation

Ensure an impartial and objective assessment, removing any potential bias or subjectivity that may be an influence in traditional grading methods.

Perfect your assignments

With the “Write with AI” tool, transform your ideas into words with a few simple clicks. Excel at all your essays, assignments, reports etc. and witness your writing skills soar to new heights

For teachers

Revolutionize your Teaching Methods

Spend less on grading

Embrace the power of efficiency and instant feedback with our cutting-edge tool, designed to save you time while providing a fair and unbiased evaluation, delivering consistent and objective feedback.

Reach out to more students

Upload documents in bulk and establish your custom assessment criteria, ensuring a tailored evaluation process. Expand your reach and impact by engaging with more students.

Focus on what you love

Let AI Grading handle the heavy lifting of assessments for you. With its data-driven algorithms and standardized criteria, it takes care of all your grading tasks, freeing up your valuable time to do what you're passionate about: teaching.

Grader Rubrics

Pick the systematic frameworks that work as guidelines for assessing and evaluating the quality, proficiency, and alignment of your work, allowing for consistent and objective grading without any bias.

Analytical Thinking

Originality

Organization

Focus Point

Write with AI

Set your tone and keywords, and generate brilliance through your words

AI Grader Average Deviation from Real Grade

Our AI grader matches human scores 82% of the time* AI Scores are 100% consistent**

Deviation from real grade (10 point scale)

Graph: A dataset of essays were graded by professional graders on a range of 1-10 and cross-referenced against the detailed criteria within the rubric to determine their real scores. Deviation was defined by the variation of scores from the real score. The graph contains an overall score (the average of all criterias) as well as each individual criteria. The criteria are the premade criteria available on Smodin's AI Grader, listed in the graph as column headings. The custom rubrics were made using Smodin's AI Grader custom criteria generator to produce each criteria listed in Smodin's premade criterias (the same criteria as the column headings). The overall score for Smodin Premade Rubrics matched human scores 73% of the time with our advanced AI, while custom rubrics generated by Smodin's custom rubric generator matched human grades 82% of the time with our advanced AI. The average deviation from the real scores for all criteria is shown above.

* Rubrics created using Smodin's AI custom criteria matched human scores 82% of the time on the advanced AI setting. Smodin's premade criteria matched human scores 73% of the time. When the AI score differed from the human scores, 86% of the time the score only differed by 1 point on a 10 point scale.

** The AI grader provides 100% consistency, meaning that the same essay will produce the same score every time it's graded. All grades used in the data were repeated 3 times and produced 100% consistency across all 3 grading attempts.

AI Feedback

Unleash the Power of Personalized Feedback: Elevate Your Writing with the Ultimate Web-based Feedback Tool

Elevate your essay writing skills with Smodin AI Grader, and achieve the success you deserve with Smodin. the ultimate AI-powered essay grader tool. Whether you are a student looking to improve your grades or a teacher looking to provide valuable feedback to your students, Smodin has got you covered. Get objective feedback to improve your essays and excel at writing like never before! Don't miss this opportunity to transform your essay-writing journey and unlock your full potential.

Smodin AI Grader: The Best AI Essay Grader for Writing Improvement

As a teacher or as a student, writing essays can be a daunting task. It takes time, effort, and a lot of attention to detail. But what if there was a tool that could make the process easier? Meet Smodin Ai Grader, the best AI essay grader on the market that provides objective feedback and helps you to improve your writing skills.

Objective Feedback with Smodin - The Best AI Essay Grader

Traditional grading methods can often be subjective, with different teachers providing vastly different grades for the same piece of writing. Smodin eliminates this problem by providing consistent and unbiased feedback, ensuring that all students are evaluated fairly. With advanced algorithms, Smodin can analyze and grade essays in real-time, providing instant feedback on strengths and weaknesses.

Improve Your Writing Skills with Smodin - The Best AI Essay Grader

Smodin can analyze essays quickly and accurately, providing detailed feedback on different aspects of your writing, including structure, grammar, vocabulary, and coherence. By identifying areas that need improvement and providing suggestions on how to make your writing more effective, if Smodin detects that your essay has a weak thesis statement, it will provide suggestions on how to improve it. If it detects that your essay has poor grammar, it will provide suggestions on how to correct the errors. This makes it easier for you to make improvements to your essay and get better grades and become a better writer.

Smodin Ai Grader for Teachers - The Best Essay Analysis Tool

For teachers, Smodin can be a valuable tool for grading essays quickly and efficiently, providing detailed feedback to students, and helping them improve their writing skills. With Smodin Ai Grader, teachers can grade essays in real-time, identify common errors, and provide suggestions on how to correct them.

Smodin Ai Grader for Students - The Best Essay Analysis Tool

For students, Smodin can be a valuable tool for improving your writing skills and getting better grades. By analyzing your essay's strengths and weaknesses, Smodin can help you identify areas that need improvement and provide suggestions on how to make your writing more effective. This can be especially useful for students who are struggling with essay writing and need extra help and guidance.

Increase your productivity - The Best AI Essay Grader

Using Smodin can save you a lot of time and effort. Instead of spending hours grading essays manually or struggling to improve your writing without feedback, you can use Smodin to get instant and objective feedback, allowing you to focus on other important tasks.

Smodin is the best AI essay grader on the market that uses advanced algorithms to provide objective feedback and help improve writing skills. With its ability to analyze essays quickly and accurately, Smodin can help students and teachers alike to achieve better results in essay writing.

© 2024 Smodin LLC

• View on Facebook Page (Opens in a new tab)
• View our Twitter Page (Opens in a new tab)
• View our Instagram Page (Opens in a new tab)
• View our Youtube Page (Opens in a new tab)

I Tested Three AI Essay-writing Tools, and Here’s What I Found

Writing essays can be draining, tedious, and difficult, even for me—and I write all day long for a living. If writing isn’t your special skill, it’s even harder, which is why there are so many sites and products out there that are designed to help you get your homework done. Some of them are pretty unethical, and I’m not going to recommend hiring someone else to write your papers for you, but there are some cool AI tools that can give you a hand that are worth considering. (The essay-writing businesspeople are probably using these, too, so you’re better off eliminating the middleman and using them on your own.)

The best AI essay-helper tools

I have an essay due next week on the history and impact of a federal law, 21 U.S.C. S856, which outlaws the operation of any building where drugs are made or used. I won’t lie: I’m excited to work on it this weekend, but that’s just me. I tested out a few of the more popular AI essay-helper tools, pretending I wasn’t excited about it, to see how they worked. Here’s my assessment. 

First up was Grammarly , which prompted me to fill out a personalization quiz before I could use it. I told the site I was a grad student, interested in improving the vocabulary I use in my work, and looking to brainstorm topics for my essay. I used the text-input section to type a quick introductory paragraph and selected “Generative AI” from the list of options. When I hit the “Improve It” button, Grammarly showed me a revised version that added a bunch of words, but still said the same thing as my more concise entry. To me, that’s annoying, but if you’re trying to hit a word count, this could be useful. I was also given options like “Make it assertive,” “Make it persuasive,” and “Make it confident.” When I selected “Make it more detailed,” the generative AI did expand the information pretty significantly, but it didn’t add any citations and I’m not convinced it drew on material outside of what I inputted. When I hit “Make it persuasive,” the AI automatically assumed the bias should fall in favor of the law, but when I added more detail to my original paragraph, suggesting for argument’s sake that the law has curtailed efforts to reduce drug overdoses throughout the country, the AI assistant said, “Grammarly assistance is unavailable for this prompt because it may result in sensitive content.”

Overall, this wasn’t great for my needs, as my topic's content was too “sensitive” and the generative AI really only added a bunch of words. This one would be most useful for someone trying to hit a word count. 

Next I tested Cramly , which I hadn’t heard of before. Before upgrading to a $4/month plan, you do get to try five free prompts, so I pasted my basic intro paragraph in there and, after a few seconds, got five paragraphs in return. It was obvious the AI was pulling from external sources somehow, as it mentioned fines and prison sentences associated with the law that I hadn’t specified, but it didn’t actually cite those sources. Still, the information it provided was helpful, so this one would be great if you’re not sure how to frame or expand on a topic and need a general idea of what your essay could look like. You’ll just have to go through everything it spits back at you and look it up independently, finding solid sourcing. 

EssayGenius

EssayGenius is extremely easy to use. It asked me to type what I’m writing about into a box. I simply inputted the name of the law and, about 10 seconds later, got 10 paragraphs back, some with subheadings like “historical background” and “implications in criminal law and public policy.” Again, there were no citations here, but the service provided not only a lot of details that could be used as a springboard to find more, but a solid outline for what the paper could look like. The AI played both sides, objectively presenting the cases for and against the law, then provided a conclusion that made it easy to narrow down where to go with the topic. Impressively, I was able to generate all that for free, but if you want to write up to 10 essays per month, it’ll cost you $9.99 a month. 

Finally, I tried out JotBot , which I have seen advertised on social media. It asked me what I was writing about, plus if I wanted an outline, but also gave me an opportunity to upload my old essays so it could replicate my writing style. As scary as it was, the paragraphs it generated after reading some of my older work did sound more like me than standard AI does. It give me subtopic suggestions, like “impact,” “historical background,” and “controversies,” which I could select from a sidebar and, if I liked the paragraph it wrote, drag into the essay itself. From there, I could accept or reject sentences one by one as it generated new ones. I could write in the essay editing section, too, and it generated more suggestions based on what I was typing. Frankly, this one was really cool and I can see how it would help beat writer’s block with ease, since you can type and get suggestions as you go. There was a learning curve, though, and I didn’t realize how quickly I was blowing through my 10 free daily “credits,” since it’s unclear what, exactly, costs credits and how many it costs to, say, accept one suggestion. You can unlock unlimited credits, unlimited autocomplete, unlimited sources, and more for $14 per month. 

Conclusions

Overall, EssayGenius and JotBot were the best AI tools I tested. I was impressed by EssayGenius’s ability to research the topic on its own and JotBot’s mimicry of my own writing style. They do cost money, but that might be worth it if you’re someone who struggles with idea generation, outline creation, or getting into the flow of writing overall. 

Bear in mind these are not meant to churn out entire essays for you and you shouldn’t use them to do that. I don’t mean because it’s unethical, but I mean because it’s pretty easy for professors to catch you doing it . Even if you do use an AI tool to generate a whole paragraph or more, try to write it in your own voice and think of it more as a way to study and learn about your topic than have the writing done for you. 

Or, pay nothing and just use ChatGPT to generate outline ideas. I do that all the time and never have to worry about getting in trouble, cheating myself out of an education, or paying for anything. I just asked ChatGPT to generate an outline for an essay on 21 U.S.C. §856 and its impact on American harm reduction efforts and got eight sections, each with three subsections, and an easy roadmap I could follow to write my paper on my own. Doing it this way ensures I’ll actually research and learn about the topic, which is important to me, but also avoid the risk of going down for plagiarism or cheating, which is probably important to you. 

No matter what you end up doing, always run your work through a plagiarism checker (like Grammarly’s , which is better than its AI essay-writing tools) and ZeroGPT to make sure you’re not turning in something that’s going to get you in trouble.

This site uses various technologies, as described in our Privacy Policy, for personalization, measuring website use/performance, and targeted advertising, which may include storing and sharing information about your site visit with third parties. By continuing to use this website you consent to our Privacy Policy and Terms of Use .

   COVID-19 Update: To help students through this crisis, The Princeton Review will continue our "Enroll with Confidence" refund policies. For full details, please click here.

We are experiencing sporadically slow performance in our online tools, which you may notice when working in your dashboard. Our team is fully engaged and actively working to improve your online experience. If you are experiencing a connectivity issue, we recommend you try again in 10-15 minutes. We will update this space when the issue is resolved.

• College Admissions Essay Counseling College Admissions Essay Counseling
• Homework Essay Feedback Homework Essay Feedback

AI Meets World-class Experts

40+ years of experience infused into the latest technology

Why The Princeton Review with AI

By leveraging AI, we can help more students more effectively and efficiently.

How We Built It

40+ years of expertise built into the latest tech. We are the next evolution of learning.

Our Vision in AI

This is just the beginning of what we aim to accomplish. Many more areas await the smartest way to achieve your goal.

We've been busy bringing these great tools into your life

Ai college admissions essay counseling, ai homework essay review.

Enrollment Advisor

1-800-2REVIEW (800-273-8439) ext. 1

1-877-LEARN-30

Mon-Fri 9AM-10PM ET

Sat-Sun 9AM-8PM ET

Student Support

1-800-2REVIEW (800-273-8439) ext. 2

Mon-Fri 9AM-9PM ET

Sat-Sun 8:30AM-5PM ET

Partnerships

• Teach or Tutor for Us

College Readiness

International

Advertising

Affiliate/Other

• Enrollment Terms & Conditions
• Accessibility
• Cigna Medical Transparency in Coverage

Register Book

Local Offices: Mon-Fri 9AM-6PM

• SAT Subject Tests

Academic Subjects

• Social Studies

Find the Right College

• College Rankings
• College Advice
• Applying to College
• Financial Aid

School & District Partnerships

• Professional Development
• Advice Articles
• Private Tutoring
• Mobile Apps
• International Offices
• Work for Us
• Affiliate Program
• Partner with Us
• Advertise with Us
• International Partnerships
• Our Guarantees
• Accessibility – Canada

Privacy Policy | CA Privacy Notice | Do Not Sell or Share My Personal Information | Your Opt-Out Rights | Terms of Use | Site Map

©2024 TPR Education IP Holdings, LLC. All Rights Reserved. The Princeton Review is not affiliated with Princeton University

TPR Education, LLC (doing business as “The Princeton Review”) is controlled by Primavera Holdings Limited, a firm owned by Chinese nationals with a principal place of business in Hong Kong, China.

Feedback Generator

AI Feedback Generator: Create Constructive and Actionable Feedback

Providing effective feedback is crucial for personal and professional growth, but crafting the right message can be challenging. 

With the AI Feedback Generator , you can create constructive and actionable feedback that helps individuals improve their performance and achieve their goals.

How to Use the AI Feedback Generator

Creating thoughtful and effective feedback with our AI Feedback Generator is a simple process. Just follow these steps:

Provide context , relevant information, and your observations by answering our thoughtful questions. This helps our AI understand the situation and create personalized feedback that addresses specific areas for improvement.

Skip questions if needed : If you don't know the answer to a question or prefer not to answer, feel free to skip it.

Review and refine : If you're not satisfied with the generated feedback, you can always go back and change your answers or add additional instructions and preferences in the last answer. Just describe your desired changes in plain English.

Craft Constructive and Actionable Feedback

Our AI Feedback Generator is designed to help you create feedback that:

Focuses on growth : Encourage personal and professional development by highlighting areas for improvement and offering actionable steps.

Balances praise and criticism : Recognize achievements and strengths while addressing areas that need attention, fostering a positive and supportive environment.

Is specific and clear : Offer precise and well-defined feedback to ensure that the individual understands the expectations and knows how to make improvements.

Empower Others with Thoughtful Feedback

Use the AI Feedback Generator to create constructive and actionable feedback that supports the growth and development of others. By providing clear and specific guidance, you can help individuals overcome challenges, enhance their performance, and achieve their goals. Get started now!

Information

• Author Services

Initiatives

You are accessing a machine-readable page. In order to be human-readable, please install an RSS reader.

All articles published by MDPI are made immediately available worldwide under an open access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. For articles published under an open access Creative Common CC BY license, any part of the article may be reused without permission provided that the original article is clearly cited. For more information, please refer to https://www.mdpi.com/openaccess .

Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.

Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive positive feedback from the reviewers.

Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world. Editors select a small number of articles recently published in the journal that they believe will be particularly interesting to readers, or important in the respective research area. The aim is to provide a snapshot of some of the most exciting work published in the various research areas of the journal.

Original Submission Date Received: .

• Active Journals
• Find a Journal
• Proceedings Series
• For Authors
• For Reviewers
• For Editors
• For Librarians
• For Publishers
• For Societies
• For Conference Organizers
• Open Access Policy
• Institutional Open Access Program
• Special Issues Guidelines
• Editorial Process
• Research and Publication Ethics
• Article Processing Charges
• Testimonials
• Preprints.org
• SciProfiles
• Encyclopedia

Article Menu

• Subscribe SciFeed
• Recommended Articles
• Author Biographies
• Google Scholar
• on Google Scholar
• Table of Contents

Find support for a specific problem in the support section of our website.

Please let us know what you think of our products and services.

Visit our dedicated information section to learn more about MDPI.

JSmol Viewer

Toward robust arabic ai-generated text detection: tackling diacritics challenges.

1. Introduction

1.1. arabic diacritized texts background, 1.2. the impact of diacritics on ai detection of arabic texts.

• Will training the classifier on diacritized texts enhance the detector’s robustness towards such texts?
• Does training the detector on diacritics-laden texts prove more effective than implementing a dediacritization filter prior to the classification step?
• Should the dataset include a diverse range of HWTs from various domains/fields to enhance recognition capabilities, or is it sufficient to focus on one domain with multiple writing styles?
• How do different transformer-based, pre-trained models (AraELECTRA, AraBERT, XLM-R, and mBERT) compare in terms of performance when trained on diacritized versus dediacritized texts?
• What is the impact of using a dediacritization filter during evaluation on the overall accuracy and robustness of the detection models across different OOD datasets?

2. Related Works

3. methodology, 3.1. data collection and preparation, 3.1.1. diacritized custom dataset, 3.1.2. ejabah-driven datasets.

• Inclusion Criteria: ⮚ Length: We included responses that exceeded 350 characters. This criterion was established because shorter responses often do not exhibit the distinct characteristics of AIGTs. For instance, a response to a simple query like “What is the capital city of the USA?” may yield a brief answer such as “Washington DC” (13 characters), which lacks the detailed features typical of LLM responses. ⮚ Balance: We ensured that the number of AIGT examples matched the number of HWT examples for a balanced comparison.
• Exclusion Criteria: ⮚ Short Responses: Responses under 350 characters were excluded as they typically fail to demonstrate the detailed and coherent characteristics of LLM outputs.
• Inclusion Criteria: ⮚ Authenticity: Only responses that were clearly identifiable as HWTs were included. If a response did not provide the source of the reply (such as books or website links), we copied the reply and searched for it on Google. This method consistently revealed that the answers were copied from online sources written by humans. ⮚ Length: Unlike AIGT samples, we accepted human responses regardless of character count because humans tend to provide shorter answers or supplement their brief answers with readily available internet text (from forums and Wikipedia), especially in community Q&A formats.
• Exclusion Criteria: ⮚ Uncertain Origin: To maintain the integrity of the HWT dataset, examples that could not be definitively classified as HWT were excluded. ⮚ Adversarial Texts: Human responses that showed signs of being a combination of AIGT and HWT, which might occur in adversarial attacks, were also excluded.

Religious Dataset

Custom plus dataset, custom max dataset, 3.2. detector architecture, 3.2.1. training, 3.2.2. evaluation, 3.2.3. hyperparameters, 3.2.4. dediacritization filter function, 3.3. evaluation strategy and methodology, 3.3.1. diacritization impact evaluation.

• Diacritized Validation set: Contains 612 examples.
• Diacritized Testing set: Contains 612 examples.
• Religious Testing set: Contains 334 examples.
• AIRABIC Benchmark Testing set: Contains 1000 examples.
• Religious Validation set: Contains 334 examples.
• Diacritized Custom Testing set: Contains 612 examples.
• Custom Testing dataset (diacritics-free): Contains 306 examples.
• AIRABIC Benchmark Testing set with dediacritization filter: The diacritics are removed from the testing set using a dediacritization filter.

3.3.2. Evaluation of Larger and More Diverse Training Corpus and Comparative Analysis of Preprocessing Approaches

Non-preprocessing approach.

• Validation set: Contains 900 examples.
• Testing set: Contains 900 examples.
• Custom dataset: This contains 306 examples.
• AIRABIC Benchmark Testing set: Contains 1000 examples used to benchmark the model’s performance against the standard dataset.
• AIRABIC Benchmark Testing set with dediacritization filter: The diacritics are removed from the testing set using a dediacritization filter to assess the model’s adaptability to diacritic-free text.

Preprocessing Approach

3.3.3. combining all datasets for maximizing the corpus size, 4.1. diacritized datasets results, 4.2. evaluation of custom plus results, 4.2.1. non-preprocessing approach, 4.2.2. preprocessing approach, 4.3. combined dataset evaluation, 5. discussion, 5.1. comparison of our detection models with gptzero on the airabic benchmark dataset, 5.2. analysis of the diacritized datasets results, 5.2.1. diacritization involvement benefit, 5.2.2. importance of diverse examples, 5.3. analysis of the results from custom plus and custom max, 5.4. takeaway recommendations for handling diacritics in arabic text detection, 5.5. in-depth analysis of models’ performance, 5.5.1. preliminary investigation of models’ tokenization and embeddings.

• Multilingual Models (mBERT and XLM-R):
• Arabic-Specific Models (AraBERT and AraELECTRA):

5.5.2. Performance Evaluation of Fine-Tuned Models

5.6. limitations, 5.7. ethical considerations, 6. conclusions and future work, author contributions, institutional review board statement, informed consent statement, data availability statement, conflicts of interest.

• Brown, T.; Mann, B.; Ryder, N.; Subbiah, M.; Kaplan, J.D.; Dhariwal, P.; Neelakantan, A.; Shyam, P.; Sastry, G.; Askell, A. Language models are few-shot learners. Adv. Neural Inf. Process. Syst. 2020 , 33 , 1877–1901. [ Google Scholar ]
• Chowdhery, A.; Narang, S.; Devlin, J.; Bosma, M.; Mishra, G.; Roberts, A.; Barham, P.; Chung, H.W.; Sutton, C.; Gehrmann, S. Palm: Scaling language modeling with pathways. arXiv 2022 , arXiv:2204.02311. [ Google Scholar ]
• OpenAI. ChatGPT (Mar 14 Version) [Large Language Model]. Available online: https://chat.openai.com/chat (accessed on 30 March 2023).
• Bard, G.A. BARD. Available online: https://bard.google.com/ (accessed on 10 October 2023).
• Gemini. Available online: https://gemini.google.com/app (accessed on 1 February 2024).
• Weidinger, L.; Uesato, J.; Rauh, M.; Griffin, C.; Huang, P.-S.; Mellor, J.; Glaese, A.; Cheng, M.; Balle, B.; Kasirzadeh, A. Taxonomy of risks posed by language models. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, Seoul, Republic of Korea, 21–24 June 2022; pp. 214–229. [ Google Scholar ]
• Sheng, E.; Chang, K.-W.; Natarajan, P.; Peng, N. Societal biases in language generation: Progress and challenges. arXiv 2021 , arXiv:2105.04054. [ Google Scholar ]
• Zhuo, T.Y.; Huang, Y.; Chen, C.; Xing, Z. Exploring ai ethics of chatgpt: A diagnostic analysis. arXiv 2023 , arXiv:2301.12867. [ Google Scholar ]
• Cotton, D.R.; Cotton, P.A.; Shipway, J.R. Chatting and cheating: Ensuring academic integrity in the era of ChatGPT. Innov. Educ. Teach. Int. 2023 , 61 , 228–239. [ Google Scholar ] [ CrossRef ]
• Gao, C.A.; Howard, F.M.; Markov, N.S.; Dyer, E.C.; Ramesh, S.; Luo, Y.; Pearson, A.T. Comparing scientific abstracts generated by ChatGPT to original abstracts using an artificial intelligence output detector, plagiarism detector, and blinded human reviewers. BioRxiv 2022 . [ Google Scholar ] [ CrossRef ]
• Anderson, N.; Belavy, D.L.; Perle, S.M.; Hendricks, S.; Hespanhol, L.; Verhagen, E.; Memon, A.R. AI did not write this manuscript, or did it? Can we trick the AI text detector into generated texts? The potential future of ChatGPT and AI in Sports & Exercise Medicine manuscript generation. BMJ Open Sport Exerc. Med. 2023 , 9 , e001568. [ Google Scholar ]
• Pegoraro, A.; Kumari, K.; Fereidooni, H.; Sadeghi, A.-R. To ChatGPT, or not to ChatGPT: That is the question! arXiv 2023 , arXiv:2304.01487. [ Google Scholar ]
• Alshammari, H.; El-Sayed, A.; Elleithy, K. Ai-generated text detector for arabic language using encoder-based transformer architecture. Big Data Cogn. Comput. 2024 , 8 , 32. [ Google Scholar ] [ CrossRef ]
• Alshammari, H.; Ahmed, E.-S. AIRABIC: Arabic Dataset for Performance Evaluation of AI Detectors. In Proceedings of the 2023 International Conference on Machine Learning and Applications (ICMLA), Jacksonville, FL, USA, 15–17 December 2023; pp. 864–870. [ Google Scholar ]
• Farghaly, A.; Shaalan, K. Arabic natural language processing: Challenges and solutions. ACM Trans. Asian Lang. Inf. Process. 2009 , 8 , 1–22. [ Google Scholar ] [ CrossRef ]
• Obeid, O.; Zalmout, N.; Khalifa, S.; Taji, D.; Oudah, M.; Alhafni, B.; Inoue, G.; Eryani, F.; Erdmann, A.; Habash, N. CAMeL tools: An open source python toolkit for Arabic natural language processing. In Proceedings of the Twelfth Language Resources and Evaluation Conference ; European Language Resources Association: Marseille, France, 2020; pp. 7022–7032. [ Google Scholar ]
• Darwish, K.; Habash, N.; Abbas, M.; Al-Khalifa, H.; Al-Natsheh, H.T.; Bouamor, H.; Bouzoubaa, K.; Cavalli-Sforza, V.; El-Beltagy, S.R.; El-Hajj, W. A panoramic survey of natural language processing in the Arab world. Commun. ACM 2021 , 64 , 72–81. [ Google Scholar ] [ CrossRef ]
• Habash, N.Y. Introduction to Arabic natural language processing. Synth. Lect. Hum. Lang. Technol. 2010 , 3 , 1–187. [ Google Scholar ]
• Abbache, M.; Abbache, A.; Xu, J.; Meziane, F.; Wen, X. The Impact of Arabic Diacritization on Word Embeddings. ACM Trans. Asian Low-Resour. Lang. Inf. Process. 2023 , 22 , 1–30. [ Google Scholar ] [ CrossRef ]
• Al-Khalifa, S.; Alhumaidhi, F.; Alotaibi, H.; Al-Khalifa, H.S. ChatGPT across Arabic Twitter: A Study of Topics, Sentiments, and Sarcasm. Data 2023 , 8 , 171. [ Google Scholar ] [ CrossRef ]
• Alshalan, R.; Al-Khalifa, H. A deep learning approach for automatic hate speech detection in the saudi twittersphere. Appl. Sci. 2020 , 10 , 8614. [ Google Scholar ] [ CrossRef ]
• El-Alami, F.-z.; El Alaoui, S.O.; Nahnahi, N.E. Contextual semantic embeddings based on fine-tuned AraBERT model for Arabic text multi-class categorization. J. King Saud Univ. Comput. Inf. Sci. 2022 , 34 , 8422–8428. [ Google Scholar ] [ CrossRef ]
• Inoue, G.; Alhafni, B.; Baimukan, N.; Bouamor, H.; Habash, N. The interplay of variant, size, and task type in Arabic pre-trained language models. arXiv 2021 , arXiv:2103.06678. [ Google Scholar ]
• Antoun, W.; Baly, F.; Hajj, H. Arabert: Transformer-based model for arabic language understanding. arXiv 2020 , arXiv:2003.00104. [ Google Scholar ]
• Antoun, W.; Baly, F.; Hajj, H. AraELECTRA: Pre-training text discriminators for Arabic language understanding. arXiv 2020 , arXiv:2012.15516. [ Google Scholar ]
• Abdelali, A.; Darwish, K.; Durrani, N.; Mubarak, H. Farasa: A fast and furious segmenter for arabic. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Demonstrations, San Diego, CA, USA, 12–17 June 2016. [ Google Scholar ]
• MADAMIRA. Available online: https://camel.abudhabi.nyu.edu/madamira/ (accessed on 7 May 2023).
• Shijaku, R.; Canhasi, E. ChatGPT Generated Text Detection. 2023; Unpublished . [ Google Scholar ]
• Guo, B.; Zhang, X.; Wang, Z.; Jiang, M.; Nie, J.; Ding, Y.; Yue, J.; Wu, Y. How close is chatgpt to human experts? comparison corpus, evaluation, and detection. arXiv 2023 , arXiv:2301.07597. [ Google Scholar ]
• Rajpurkar, P.; Zhang, J.; Lopyrev, K.; Liang, P. Squad: 100,000+ questions for machine comprehension of text. arXiv 2016 , arXiv:1606.05250. [ Google Scholar ]
• Fan, A.; Jernite, Y.; Perez, E.; Grangier, D.; Weston, J.; Auli, M. ELI5: Long form question answering. arXiv 2019 , arXiv:1907.09190. [ Google Scholar ]
• Liu, Y.; Ott, M.; Goyal, N.; Du, J.; Joshi, M.; Chen, D.; Levy, O.; Lewis, M.; Zettlemoyer, L.; Stoyanov, V. Roberta: A robustly optimized bert pretraining approach. arXiv 2019 , arXiv:1907.11692. [ Google Scholar ]
• Antoun, W.; Mouilleron, V.; Sagot, B.; Seddah, D. Towards a Robust Detection of Language Model Generated Text: Is ChatGPT that Easy to Detect? arXiv 2023 , arXiv:2306.05871. [ Google Scholar ]
• Antoun, W.; Sagot, B.; Seddah, D. Data-Efficient French Language Modeling with CamemBERTa. arXiv 2023 , arXiv:2306.01497. [ Google Scholar ]
• Martin, L.; Muller, B.; Suárez, P.J.O.; Dupont, Y.; Romary, L.; de La Clergerie, É.V.; Seddah, D.; Sagot, B. CamemBERT: A tasty French language model. arXiv 2019 , arXiv:1911.03894. [ Google Scholar ]
• Clark, K.; Luong, M.-T.; Le, Q.V.; Manning, C.D. Electra: Pre-training text encoders as discriminators rather than generators. arXiv 2020 , arXiv:2003.10555. [ Google Scholar ]
• Conneau, A.; Khandelwal, K.; Goyal, N.; Chaudhary, V.; Wenzek, G.; Guzmán, F.; Grave, E.; Ott, M.; Zettlemoyer, L.; Stoyanov, V. Unsupervised cross-lingual representation learning at scale. arXiv 2019 , arXiv:1911.02116. [ Google Scholar ]
• Zaitsu, W.; Jin, M. Distinguishing ChatGPT (-3.5,-4)-generated and human-written papers through Japanese stylometric analysis. PLoS ONE 2023 , 18 , e0288453. [ Google Scholar ] [ CrossRef ]
• GPTZero. Available online: https://gptzero.me/ (accessed on 1 June 2023).
• OpenAI. Available online: https://beta.openai.com/ai-text-classifier (accessed on 1 June 2023).
• Schaaff, K.; Schlippe, T.; Mindner, L. Classification of Human-and AI-Generated Texts for English, French, German, and Spanish. arXiv 2023 , arXiv:2312.04882. [ Google Scholar ]
• Devlin, J.; Chang, M.-W.; Lee, K.; Toutanova, K. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv 2018 , arXiv:1810.04805. [ Google Scholar ]
• Loshchilov, I.; Hutter, F. Decoupled weight decay regularization. arXiv 2017 , arXiv:1711.05101. [ Google Scholar ]
• Yafooz, W.M. Enhancing Arabic Dialect Detection on Social Media: A Hybrid Model with an Attention Mechanism. Information 2024 , 15 , 316. [ Google Scholar ] [ CrossRef ]
• Aldyaflah, I.M.; Zhao, W.; Yang, S.; Luo, X. The Impact of Input Types on Smart Contract Vulnerability Detection Performance Based on Deep Learning: A Preliminary Study. Information 2024 , 15 , 302. [ Google Scholar ] [ CrossRef ]
• Khalif, K.M.N.K.; Chaw Seng, W.; Gegov, A.; Bakar, A.S.A.; Shahrul, N.A. Integrated Generative Adversarial Networks and Deep Convolutional Neural Networks for Image Data Classification: A Case Study for COVID-19. Information 2024 , 15 , 58. [ Google Scholar ] [ CrossRef ]
• Pardede, H.F.; Adhi, P.; Zilvan, V.; Yuliani, A.R.; Arisal, A. A generalization of sigmoid loss function using tsallis statistics for binary classification. Neural Process. Lett. 2023 , 55 , 5193–5214. [ Google Scholar ] [ CrossRef ]

Click here to enlarge figure

Share and Cite

Alshammari, H.; Elleithy, K. Toward Robust Arabic AI-Generated Text Detection: Tackling Diacritics Challenges. Information 2024 , 15 , 419. https://doi.org/10.3390/info15070419

Alshammari H, Elleithy K. Toward Robust Arabic AI-Generated Text Detection: Tackling Diacritics Challenges. Information . 2024; 15(7):419. https://doi.org/10.3390/info15070419

Alshammari, Hamed, and Khaled Elleithy. 2024. "Toward Robust Arabic AI-Generated Text Detection: Tackling Diacritics Challenges" Information 15, no. 7: 419. https://doi.org/10.3390/info15070419

Article Metrics

Article access statistics, further information, mdpi initiatives, follow mdpi.

Subscribe to receive issue release notifications and newsletters from MDPI journals