How does Learnaway work? Behavioural AI detection explained
Learnaway works by recording the writing process, not the finished prose. A teacher creates an assignment link; students write in Learnaway's editor, which captures objective behavioural signals — how much text arrived via paste, inter-keystroke timing variance, tab-switch events, and total session duration. Submissions are scored against concrete thresholds: a paste share above 60–90% of the final word count, inter-keystroke variance below 150ms (consistent with transcribing AI output), or a 1,000-word essay completed in under 8 minutes. Those process-based signals are shown on the teacher dashboard alongside the final text, giving educators defensible facts rather than an opaque probability score.
Step 1 — Create an assignment link
The teacher logs in and creates an assignment in under a minute: a title, optional instructions, and a deadline. Learnaway generates a unique link the teacher shares with students — no student accounts, no app installation, and no LMS integration required. Any channel works: email, Google Classroom, a class message.
Step 2 — Students write in the Learnaway editor
The student opens the link and types directly into Learnaway's editor. In the background, Learnaway records behavioural telemetry: each keystroke event and timing delta, paste events (character count only, never clipboard content), idle gaps, and browser tab-switch events. No characters typed are stored — only the anonymised timing metadata and the final submitted text.
Step 3 — Learnaway scores the submission
On submission, the behavioural timeline is processed by Learnaway's deterministic rule engine. Signals that raise concern include: paste share above 60% of the final word count, inter-keystroke variance below 150ms across a sustained typing run, and a session under 8 minutes for a 1,000-word assignment. Multiple signals together elevate the risk level; no single signal is treated as conclusive on its own.
Step 4 — Teacher reviews signals on the dashboard
The teacher sees a risk label (Likely Authentic, Low Concern, Elevated Concern, or High Concern) alongside the specific data points that drove it — for example, '82% of text arrived via paste; session lasted 5 minutes for a 950-word essay'. Those concrete facts support a professional conversation with the student rather than acting as a machine verdict.
FAQ
- Does Learnaway need to integrate with our LMS?
- No. Learnaway works via a plain link the teacher shares by any channel. No LMS integration, student accounts, or app installation are required.
- What writing data does Learnaway store about students?
- Timing metadata only — event types, timing deltas, paste lengths — never the characters students type or clipboard content. By default the detailed writing timeline is scored and then discarded; only the final text and risk analysis are retained.
- Can students edit their work after submitting?
- No. The editor locks on submission. This prevents post-submission changes and preserves the writing-process evidence for the teacher's review.
Want to see this on real writing?