Kritik AI Authorship Solutions
Improving AI Use in Higher Education Through Transparent Peer Assessment
Summary:
Kritik’s authorship-tracking platform gives students and instructors visibility into submissions, helping prevent false plagiarism and AI-use claims. As the sole designer, I shaped an end-to-end experience by consolidating the product suite, integrating peer assessment, and making the writing interface more accessible.
Problem: Students beginning university increasingly rely on AI for everything from everyday tasks to schoolwork, and instructors have noticed that this has led to more plagiarism and weaker critical thinking. This increases instructor workloads.
Approach: I iterated on early designs, using AI to accelerate discovery and challenge my design assumptions. I based the work on student and instructor feedback workshops to identify the highest-priority writing standards. I presented an improved editing and assessment experience to stakeholders, refining interactions to keep student learning at the center.
Impact: Increased instructors’ time-to-value by 10% & streamlined design with an agentic AI strategy to accelerate discovery and refine solutions.
Company
Kritik Education
Year
2025-Present
Role
Product Designer
Opportunity Space
We found that as AI became more common in student writing, many professors were pushed to use detection tools. We saw the potential in consolidating this AI literacy tool with our existing peer assessment toolkit and create a space for process tracking, where students could use AI openly. Peer-to-peer accountability was a built-in opportunity to provide students with a more comfortable way to learn and stay accountable.
This consolidation posed an opportunity to design a space that instructors and students could share across every stage of the peer assessment lifecycle: student creation, peer evaluation, and peer feedback.
Research Practices
I dove directly into market validation among our user base.
20+ user interviews and walkthroughs with instructors and students
Conducted a UX Audit of the existing product as a form of benchmarking
The interviews helped showcase what changes users want, which helped stakeholders align on design decisions.
Key Insights
We learned two main items from our interviews:
Students were experiencing anxiety over AI in their coursework as university policies became increasingly strict.
Students wanted clear, measurable insight into how they use AI
Instructors wanted students to take a more holistic approach to the peer assessment process to establish accountability
Key Takeaways
Users at varying levels had different goals when using this page:
Instructors are looking to grade accurately and efficiently with minimal navigation.
Students want to be able to reference materials during their writing, use the AI Assistant, and view their authorship stats.
We learned that regardless of the persona, all users wanted one focused workspace, and the challenge was going to be cleanly including all information in that one space, reducing the orginal layout:
Original Layout
Evaluate Stage
Feedback Stage
Concepts and Challenges - Kritik Writing Editor & Peer Assessment Stages
During the initial concept phase, I built generative AI critique personas to explore aesthetic variations. As we refined the designs, we asked:
A) What actions do users take here, and at what stage in the writing process?
B) What information do users reference most often?
The top two tasks were viewing the activity instructions and referencing the assignment rubric. That required extra attention to affordance across each stage, and the biggest challenge was maintaining visual consistency while supporting dynamic interactions across user views.
Create Stage
Familiarity, Accessibility, and Measurable Insights
Because one of our goals was to create a space where students can use AI openly while viewing metrics shared with their instructor, we focused on making the writing space feel familiar to them.
An Insights ribbon, and the formatting toolbar are always in view along with:
The activity title.
Peer assessment stage and due date.
The save state.
Solutions & Outcomes
Peer Evaluation and Accountability
Students are provided clear guidance while evaluating peer work, with the rubric visible beside each submission.
This helps them provide fair, specific, and constructive feedback while tracking how their peers used AI to shape their work. Students can also:
View their peers’ entire AI chat history
Track the number of assigned and completed evaluations.
Celebrate progress when an evaluation is finished.
Solutions & Outcomes
Solutions & Outcomes
Feedback on Feedback
In the Kritik platform, students provide feedback on the evaluations they received.
This gives them the opportunity to hone their AI Literacy skills by being able to:
Review their AI chat history and writing process
Compare their submission against their peers’ thoughts and comments
Results and Impact
Students and Instructors were overall extremely pleased with the aesthetics and visibility brought to the product while integrating the two toolkits, and as a team we were able to successfully establish an AI assisted design workflow that provided quick turnaround and improved collaboration with the engineering team.
Functionally, instructors described the new layout and workflow as “where we want to be headed“ and a pedagogical success. For example, the mental model for users navigating the peer assessment workflow moved from being a step-by-step process to a holistic and consistent overview and review of a peer-assessed assignment. Students found the writing experience similar to Google Docs. This resulted in a visible change in how users approach their assignments in Kritik.
Project Lessons
At the outset of this project we saved three AI critique prompts as reusable skills, each reviewing a design as a different person: a professor, a product designer, and a student. Any screenshot or Figma frame could then get a structured review in minutes. The professor critique asked what an instructor needs to defend a grade, such as which student an AI-use score belongs to, while the designer critique removed clutter. Where they agreed, we made those changes first.
Always design with a combination of raw data and human conversations, particularly in an AI-accelerated workflow. We learned that the critiques are only as good as what we tell them. The professor critique flagged an AI-use limit as confusing, even though professors set that limit themselves. It also suggested making an instructor Comment button more prominent, but our company data showed only 59 of about 2.43 million student-to-student evaluations ever got an instructor comment. We learned to save each correction as a rule, and treat the critiques as a fast first filter, not a replacement for speaking with users.