Gradescope

What is Gradescope AI?

Gradescope is an academic assessment and grading platform founded in 2014 by Sergey Karayev and Arjun Singh at UC Berkeley. Acquired by Turnitin in 2018, the tool uses computer vision and artificial intelligence to assist teachers in grading handwritten exams, bubblesheets, online homework, and computer programming projects, reducing grading time and improving evaluation consistency.

How it works

The instructor uploads an exam template and creates a grading rubric with specific point values. Students upload their completed exams (as photos or PDFs) or teachers scan paper exams in batches. Gradescope's AI reads the handwriting and automatically groups similar answers together. Instead of grading one student's entire exam at a time, the teacher grades all answers to a specific question grouped together by the AI (e.g., grading all identical answers to Question 1 in one click).

Key features

• AI answer grouping: Automatically groups identical or highly similar answers for multiple-choice, math, or short text questions.
• Dynamic rubrics: Change point values during grading and retroactively apply the new values to all previously graded exams.
• Programming Autograder: Runs unit tests on student code submissions to assign scores automatically.
• Detailed statistics: Metrics showing class performance on specific concepts or questions to identify learning gaps.
• Handwritten text recognition: Advanced OCR capable of reading student handwriting.

Available integrations

Gradescope integrates with popular Learning Management Systems (LMS) such as Canvas, Blackboard, Moodle, Sakai, and Brightspace, allowing roster sync and grade export.

Who it is for

• University professors and TAs managing large classes where grading takes days.
• K-12 teachers seeking a consistent way to grade math and science homework.
• Computer Science instructors looking for automated coding grading tools.

Real use cases

1. Calculus Exam Grading: An engineering professor scans 300 paper exams. Gradescope groups all Question 1 answers into five groups based on the final visual result. The professor grades each group once, saving 70% of grading time.
2. Coding Submissions Evaluation: A university uses the Autograder module to run unit tests on hundreds of Python code files submitted by students, assigning grades instantly.

Pricing

Pros and cons

Pros:

• Massive time savings in grading paper exams.
• Ensures total consistency in applying rubric points.
• Excellent support for CS code grading.

Cons:

• Initial process of scanning paper exams can be labor-intensive without high-speed scanners.
• Handwriting recognition may struggle with highly illegible writing.

Alternatives to Gradescope AI

Key competing academic grading tools are:

• Canvas SpeedGrader: Built-in LMS grader, though lacking AI grouping features.
• GradeCam: Scan sheets using mobile devices.
• Crowdmark: Collaborative grading platform for higher education.