{"stable_id":"d8b354365bb9e260","slug":"be-my-ai-visual-assistance","company":"Be My Eyes","workflow_name":"Conversational multimodal visual assistance through Be My AI","function_code":"engineering","pattern_codes":["creator_to_judge"],"changed_assumption":"A blind or low-vision user can get immediate, private scene descriptions before choosing human assistance.","evidence_strength":"verified","publication_tier":"showcase","freshness":"watch","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"1 peer reviewed; publication outcomes are verified.","caveat_summary":"Qualitative study; no causal time-saving or safety metric. Small purposive sample; limitations may change with model updates.","freshness_caveat":"Watch status: verify the cited source and deployment condition before reusing this case.","workflow_summary":{"before":[{"actor":"Blind or low-vision user","action":"Contacts a sighted volunteer or trusted person and shares a camera view.","actor_type":"control"},{"actor":"Human assistant","action":"Describes the visual scene and answers follow-up questions.","actor_type":"human"}],"hinge":"A blind or low-vision user can get immediate, private scene descriptions before choosing human assistance.","after":[{"actor":"User","action":"Captures an image and asks Be My AI for a description or follow-up detail.","actor_type":"system"},{"actor":"User","action":"Evaluates the answer and either acts, reprompts, or escalates to a human for high-stakes or uncertain tasks.","actor_type":"system"}],"decision_mode":"shared","decision_marker":"Human moves from creator to judge"},"before":[{"order":1,"actor":"Blind or low-vision user","action":"Contacts a sighted volunteer or trusted person and shares a camera view.","handoff_to":"Human assistant","control":"User explains intent and decides whether to act."},{"order":2,"actor":"Human assistant","action":"Describes the visual scene and answers follow-up questions.","handoff_to":"User","control":"User judgment."}],"after":[{"order":1,"actor":"User","action":"Captures an image and asks Be My AI for a description or follow-up detail.","handoff_to":"Multimodal model","control":"User controls framing, prompting, and use."},{"order":2,"actor":"User","action":"Evaluates the answer and either acts, reprompts, or escalates to a human for high-stakes or uncertain tasks.","handoff_to":"Action or human assistant","control":"Human escalation remains available."}],"decision_rights":"The user decides whether to trust, reprompt, or escalate; humans retain action decisions.","exception_path":"Users compensate for hallucinations or misunderstood context and intent through spatial memory, auditory cues, personal judgment, reprompting, orientation and mobility skills, or human assistance.","removed_work":["Routine requests to sighted volunteers for many household, spatial, and social descriptions"],"outcomes":[{"metric":"Demonstrated task scope in user study","baseline":"Visual interpretation commonly required sighted assistance.","result":"Fourteen users described Be My AI support across personal and social contexts.","period":"CHI 2025 qualitative study.","scale":"14 visually impaired Be My AI users plus analysis of image descriptions.","attribution_caveat":"Qualitative study; no causal time-saving or safety metric.","evidence_label":"verified"},{"metric":"Observed failure modes","baseline":"Human assistants can clarify intent and context interactively.","result":"The study identified hallucinations, misinterpretations of social contexts, styles, and human identities, and failures to grasp or act on user intent.","period":"CHI 2025 study.","scale":"14 interviews plus image descriptions from participants and social media.","attribution_caveat":"Small purposive sample; limitations may change with model updates.","evidence_label":"verified"}],"executive_lesson":"Accessibility value comes from giving the user control over when AI is sufficient and when a human is needed.","anti_pattern":"Positioning probabilistic image description as safe navigation or identity verification.","questions_for_leaders":["Which tasks are explicitly out of scope?","How quickly can users reach a human?","How are hallucinations communicated accessibly?"],"collections":["human-still-decides","negative-results"],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/be-my-ai-visual-assistance"}