{"stable_id":"33e321501f83de63","slug":"google-project-relate-accessibility","company":"Google Research Project Relate","workflow_name":"Personalized speech recognition for transcription, restatement, and text entry","function_code":"customer_service","pattern_codes":["creator_to_judge"],"changed_assumption":"People with non-standard speech need not rely only on listener familiarity, repetition, or text entry.","evidence_strength":"verified","publication_tier":"showcase","freshness":"watch","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"1 peer reviewed, 2 primary; publication outcomes are verified.","caveat_summary":"The study is a small feasibility evaluation, not proof of universal accuracy or communication benefit; Google currently accepts no new users.","freshness_caveat":"Watch status: verify the cited source and deployment condition before reusing this case.","workflow_summary":{"before":[{"actor":"Person with non-standard speech","action":"Speak, repeat, gesture, or switch to typing when a listener or device does not understand.","actor_type":"control"},{"actor":"Listener","action":"Interpret, ask for clarification, or fail to complete the interaction.","actor_type":"control"}],"hinge":"People with non-standard speech need not rely only on listener familiarity, repetition, or text entry.","after":[{"actor":"User","action":"Records at least 500 phrases to create a personalized recognition model.","actor_type":"system"},{"actor":"Project Relate","action":"Transcribes speech, restates it in a synthesized voice, or sends text through a keyboard.","actor_type":"system"},{"actor":"User and listener","action":"Use the output with active listening, gestures, and clarification.","actor_type":"system"}],"decision_mode":"shared","decision_marker":"Human moves from creator to judge"},"before":[{"order":1,"actor":"Person with non-standard speech","action":"Speak, repeat, gesture, or switch to typing when a listener or device does not understand.","handoff_to":"Listener or app","control":"Human accommodation"},{"order":2,"actor":"Listener","action":"Interpret, ask for clarification, or fail to complete the interaction.","handoff_to":"Conversation","control":"Listener understanding"}],"after":[{"order":1,"actor":"User","action":"Records at least 500 phrases to create a personalized recognition model.","handoff_to":"Project Relate","control":"Android, connectivity, and voice stability"},{"order":2,"actor":"Project Relate","action":"Transcribes speech, restates it in a synthesized voice, or sends text through a keyboard.","handoff_to":"Listener or application","control":"Personalized model"},{"order":3,"actor":"User and listener","action":"Use the output with active listening, gestures, and clarification.","handoff_to":"Conversation or workflow","control":"Humans retain meaning and communication control"}],"decision_rights":"The model proposes a transcript or synthesized restatement; the user decides whether it expresses the intended message.","exception_path":"Users repeat, type, gesture, or ask a listener for clarification when recognition is low; severe or rapidly changing speech can make the model unsuitable.","removed_work":["Some repeated speech repair","Some manual typing for supported users"],"outcomes":[{"metric":"Feasibility of personalized ASR use in Ghana","baseline":"General ASR can have 78-89% word-error rates for severely dysarthric speech, with 0-1.2% correctly transcribed sentences in cited prior work","result":"Twenty people with communication difficulties completed a Ghana user study; the study identified feasibility but also access, adaptation, support, and accuracy limitations","period":"Study published at CHI 2024","scale":"20 users plus locally trained speech and language therapists in Ghana","attribution_caveat":"The study is a small feasibility evaluation, not proof of universal accuracy or communication benefit; Google currently accepts no new users.","evidence_label":"verified"}],"executive_lesson":"Assistive AI is not useful merely because it recognizes more speech; onboarding burden, device access, changing voices, and user control determine whether it increases autonomy.","anti_pattern":"Treating a transcript as the user's intended message or presenting a closed beta as broadly available.","questions_for_leaders":["Can the user easily reject or correct the output?","What happens as the user's voice changes?"],"collections":[],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/google-project-relate-accessibility"}