Is the TCF Canada marked by AI?
Checked November 1, 2026
Partly. Since October 2024, France Éducation international (FEI) has described its results as coming from a hybrid double correction. One correction is always human, and the other is done by an automated process called FIDELIA. When the two differ significantly, a human does a third correction. FEI states that results never rest on automated processing alone.
What that means for you.
- A person always marks your work. The automated correction is only one of the two.
- FEI describes writing marking as double-blind. Each corrector gives each of the 3 tasks a level from “A1 non atteint” to C2 without seeing the other’s mark. For speaking, a second corrector listens to the recording and adds to the examiner at the centre.
- In computer-based writing tests, the manual says the first correction may be done by FIDELIA and the second is always human. If you would rather avoid that, ask to sit on paper.
- The manual does not explain how FIDELIA works or how its marks are weighed, so read what FEI does say in its manual.
NCLC Prep’s estimate is automated too, but it is a practice tool. For a text you write, it gives a score out of 20 and an NCLC range. It is not the official score.
How marking, results and complaints fit together is explained in TCF Canada results and validity.