FAQ
Clear answers before a demonstration.
What Sign First is, what it is not, what works today, and what still requires evidence.
11 / The boundaries
Trust starts by answering the difficult questions without marketing around them.
Is this an ASL translator?
No. It is an academic assurance layer.
Sign First can accept candidates from recognition systems, but it does not claim to solve universal ASL translation or replace interpreters.
Is it ready for children?
The technical research build is demonstrable; student outcomes are not validated.
Real student use requires Deaf-led product governance, school and family authorization, signer-independent evidence, and age-specific usability work.
Does use train the model?
No. Private use is the default.
Personalized support and contribution are separate choices. Raw video is not retained by the current runtime, and real visual-model contribution is disabled.
Questions people should ask
No fine print required.
Is Sign First an ASL translator?+
No. It is an academic assurance layer that can receive recognition candidates, focus review on consequential uncertainty, and require signer-controlled confirmation before AI answers.
Does it replace an interpreter?+
No. It does not replace interpreters, educators, language access, or human relationships. It addresses a narrower interface problem: signed interaction with AI.
What can the demonstration prove today?+
It can prove the implemented control path, authentication boundary, provider routing, synthetic learning-governance machinery, and evidence reporting. It cannot prove reliable student outcomes or universal ASL recognition.
Is the product only fingerspelling?+
No. The browser-local recognizer is a bounded fingerspelling specialist used to demonstrate one input path. The architecture also accepts continuous-sign candidates, but current public research candidates have not passed the release gate for direct student use.
Why does the learner confirm?+
Because fluent text can hide a meaning-changing error. Full confirmation is the current safety scaffold. The intended future is selective, confidence-calibrated intervention—not a permanent extra step on every question.
Does ordinary use train a model on a child?+
No. Private use is the default. Personalized support and voluntary model contribution are separate modes. Real visual-ASL contribution is disabled until separate governance and authorization exist.
Is raw camera video stored?+
No. The current product runtime processes the camera stream for the immediate interaction and does not retain raw video or audio.
Which AI models are connected?+
The controlled answer route demonstrates Google Vertex Gemini. Other providers are architectural options and are not described as live until they are implemented and independently tested.
Is the site ready for a school pilot?+
Not yet. A stakeholder demo is available. Any real learner pilot requires Deaf-led governance, signer-independent evidence, school and family authorization, legal review, accessibility validation, and age-specific usability work.
Continue exploring