FluencyAI LogoFluencyAI

The Arkenza Fluency Program and the Fluent Digital Twin

New uses for Artificial Intelligence and other emerging technologies

Arkenza

  • Arkenza is a proposed fluency program that expands the range of conversational situations in which people who stutter can remain fluent.
  • Users begin 'speaking while alone.'
  • Arkenza uses AI-assisted conversations to gradually introduce realism into conversations.
  • Software components and some testing expected by fall 2025.

Objectives of Arkenza

  • Arkenza is designed to enable people who stutter (PWS) to remain fluent over a wider range of conversational situations.
  • It has been known for more than 80 years that PWS are fluent when they are alone. This intriguing feature of stuttering has been confirmed in multiple research studies.
  • The expectation of disfluent speech (Anticipatory Struggle Hypothesis) increases stuttering. By removing that anticipation, we expect fluency to emerge.

Removing the Anticipation of Stuttering

  • Arkenza starts users in a speaking environment where they should experience fluency: reciting text while alone in unison with a choral reader. Through a series of graduated speaking exercises, we expect that users will gradually anticipate fluency, leading to confident speech in the presence of others.
  • Arkenza uses AI technologies, such as speech-to-text, text-to-speech, chatbot conversations, synthetic choral readers, virtual reality audiences, avatars, sanitized text passages and video conferencing, to gradually introduce more realistic conversation settings.

The Fluent Digital Twin (FDT)

A Novel Use of Technology to Create Fluent Presentations

  • A common refrain in the posts of the Reddit stuttering subgroup is the fear experienced by students and employees who stutter when they are asked to make a presentation to their peers.
  • In this post-Covid era, remote presentations by video conference calls such as Zoom, Google Meet, and Microsoft Teams, are the norm.
  • Straightforward AI processing of users' transcribed speech can remove all disfluencies. The fluent transcription can easily be reconverted back into synthetic speech by text-to- speech apps.
  • This allows stuttering users to maintain a fluent 'presence' in video conference calls: a 'Fluent Digital Twin' of the speakers.
  • To obtain an 'alpha-release' of the Fluent Digital Twin for use with Google Meet (experimental, but free), contact steve@fluencyai.net.

Contact Us

Contact Steven Scott, president, fluencyAI LLC
Email: steve@fluencyai.net