Read my lips: The importance of the face in a computer-animated tutor for vocabulary learning by children with autism.

The team built a computer tutor named Baldi. Baldi shows an animated face that moves its lips while it speaks.

Kids with autism heard new words either with the face on screen or with voice only. The study switched the two conditions back-to-back so each child served as their own control.

Children learned the new words faster when the animated face was present. They also remembered the words better the next day.

Voice-only lessons worked, but the face gave a clear boost.

Rice et al. (2015) and IFaso et al. (2016) later used animated faces too. They moved from vocabulary to emotion recognition and still saw gains, showing the idea stretches beyond word learning.

Iarocci et al. (2010) seems to disagree. They found autistic kids gain little from lip-reading cues during noisy speech. The tasks differ, though. Grace tested speech-in-noise perception; W et al. taught vocabulary on a quiet computer. When the goal is clear teaching, the face helps.

Bloh et al. (2025) used the same back-and-forth design. They compared animated videos to real-person videos and saw mixed results. Their data echo W et al.: animated models can win, but individual preference matters.

If you run computer lessons, turn on the face. A simple animated mouth can speed up vocabulary growth without extra staff time. Try it next session: play your usual audio lesson on half the trials and add a lip-sync face on the other half. Track which set reaches mastery first.