Ethics and Limitations

ASBOS must be framed as a decision-support and workflow-automation system, not as a replacement for human accountability.

Ethics

• Disclose automation during voice calls.
• Do not collect unnecessary location, biometric, or voice data.
• Preserve human override for safety-sensitive decisions.
• Avoid punitive use of raw analytics without context.
• Provide employees and managers clear explanations of decisions.
• Client-health signals must not use protected characteristics.

Current limitations

• Conceptual and design contribution only — no completed ASBOS deployment, causal effect, measured fill time, measured administrative savings, or validated ROI.
• This repository uses synthetic data only; voice outreach is simulated by default.
• The 3–6× ROI claim is withdrawn; the paper presents a transparent sensitivity model only (base case 1.35× benefit-cost ratio is illustrative, not measured).
• Market-size figures are third-party forecasts and differ by publisher and category definition.
• External labor-law rules are not fully implemented; multi-tenant production security is outside the scope of this research scaffold.
• Public product documentation may not reveal every feature, implementation constraint, or roadmap item.
• The citation search was extensive but cannot prove nonexistence of unpublished or proprietary systems.

Evidence grading (from verification method)

Future work

• Registered pilot protocol
• De-identified event dataset
• Baseline work-sampling study
• Failure-mode and effects analysis
• Fairness testing of replacement ranking
• Voice-agent usability testing across accents and noisy environments
• Independent security and privacy review