The rise of AI in education has disrupted traditional learning and assessment methods, pushing schools and educators to rethink their approach to academic integrity and student learning. While Surveillance Tech and Detection Tech have dominated the conversation, they both fall short in the AI era.

A new paradigm is emerging—Drafting Tech. This category of tools doesn’t just evaluate the final product but provides visibility into the entire writing process. It offers a more nuanced, transparent, and student-friendly way to assess originality, improve feedback, and support responsible AI usage.

Understanding the Three Approaches to AI and Academic Integrity

‍

‍

1. Surveillance Tech

Surveillance technology monitors students during assessments, using tools like browser locking, screen recording, and webcam monitoring.

• Examples: Honorlock, Proctorio
• When It Works: Standardized, high-stakes exams (e.g., TOEFL, SAT, GRE) where real-time supervision is necessary.
• Why It Fails Elsewhere: Most assignments and finals are not one-shot exams. Students typically write over multiple days, collaborate, and revise. Furthermore, surveillance creates a hostile environment—being "watched" leads to anxiety and raises ethical concerns about privacy.

2. Detection Tech

Detection tools analyze completed work to assess its originality, often using AI models or databases to identify copied content.

• Examples: Turnitin, CopyLeaks
• When It Worked: Before generative AI, when plagiarism mostly meant copy-pasting from the internet.
• Why It Fails Now:
  • AI-generated text is difficult to detect reliably. Studies, such as this research and findings from University of Maryland researchers show, that AI detection is flawed.
  • The top 20 US Universities advise against AI detectors, recognizing their inaccuracy and potential for false accusations.
  • If educators want students to use AI, these tools provide no insight into how it was used—whether properly, ethically, or effectively.

3. Drafting Tech

Drafting Tech captures every stage of the writing process, allowing educators to see how a piece of work evolved rather than just analyzing the final product. There is growing scientific evidence that keystrokes are the most reliable method to identify original work.

• Examples: Google Docs, Rumi
• How It Works:
  • Collects all drafts during the writing process.
  • Provides analytics on how students develop their ideas and whether they use AI responsibly.
  • Enables students to restore any version of their work, avoiding the limitations of periodic auto-saves.

The Benefits of 'Drafting Tech'

• Transparency Over Originality
  • Instead of trying to "catch" AI-generated work, instructors can see exactly how a piece of writing was developed.
  • Helps identify AI misuse vs. responsible use.
• Improved Feedback for Students
  • Educators can coach students on their process, not just their final product.
  • Provides insights into revision habits, time spent on tasks, and where students might struggle.
• Empowering, Not Policing, Students
  • Unlike Surveillance or Detection Tech, Drafting Tech encourages students to take ownership of their learning.
  • Students can use AI tools responsibly while still demonstrating their originality.
    ‍

Why Rumi vs. Google Docs?

‍

Google Docs tracks every keystroke and AI prompts of users but doesn't openly share this with users and 3rd party developers. Instead, some engineers have reverse-engineered Google Docs to retrieve keystrokes using backdoor methods, but this is neither transparent nor student-friendly.

Rumi also tracks all drafts and keystrokes. However, Rumi is intentionally designed to share all  collected data points with instructors and students in an open, transparent way to enhance learning and encourage ethical AI use.

Below we've highlighted the key difference between Rumi and Google Docs.

Rumi vs Google Docs

Features	Rumi	Google Docs
Real-time Group Collaboration
Seamlessly embeds within any system or LMS with SSO (Single Sign-on)
Customizable writing environment with configurable editor settings for each assignment
Instructors can pre-train the student facing AI for any assignment
Native AI integration with any LLM, including third-party models schools build, pay for, or deploy
Secure draft retrieval via supported/official APIs
Customizable document sharing and access control
Advanced protection against circumvention methods (typing simulators, unauthorized extensions)
Students can submit work directly to Canvas or any LMS
Better integration with Google Ecosystem

The Future of AI & Academic Integrity in EdTech

As AI becomes a fundamental part of education, the question isn’t whether students should use AI—it’s how they can use it ethically and effectively. Surveillance and detection tools create hostile learning environments, whereas Drafting Tech enables a culture of transparency, learning, and responsible AI use.

For educators and institutions looking to embrace AI rather than fear it, Drafting Tech is the key to the future.