AMINA: The Inclusive and Accountable AI for Marginalized Immigrant Nonprofit Assistance

URL SCAN: AMINA: The Inclusive and Accountable AI for Marginalized Immigrant Nonprofit Assistance
FIRST LINE: Computer Science > Computers and Society

The Dissection

The paper converts political exclusion into a product-design problem. It documents real barriers—registry exclusion, capacity gaps, misinformation, and dispersed networks—then frames AI as an inclusive, accountable collaborator that can make nonprofit operations more legible and efficient.

Its evidence supports prototype usefulness and user-perceived potential, not durable inclusion, scalable misinformation control, or actual accountability. The paper’s strongest operation is legitimacy work: making AI acceptable inside vulnerable communities by presenting it as relational infrastructure rather than automation.

The Core Fallacy

It confuses preserving human relationships with preserving human economic necessity.

An AI that handles reporting, reminders, information triage, and coordination may preserve the nonprofit’s social surface while automating its administrative labor. Under the Discontinuity Thesis, AI need not eliminate every human interaction. It only needs to outperform humans across enough cognitive work to make much of that labor economically unnecessary.

“Strengthens rather than displaces” is therefore not demonstrated by preserving human connections. It is a non-displacement narrative attached to a system that may reduce staffing needs and centralize gatekeeping. Transparency is also treated as if it automatically produces accountability. Records and reminders create legibility; accountability requires ownership, contestability, and power to correct the system.

Hidden Assumptions

• Better interfaces can overcome barriers rooted in political power, legal status, institutional exclusion, and platform control.
• The system will have safe data, reliable funding, human oversight, and trustworthy deployment conditions.
• AI can identify and counter politically charged misinformation without reproducing bias, surveillance, or political capture.
• Reducing reporting burdens will strengthen organizations rather than justify staff cuts or intensify remaining work.
• Multilingual access equals meaningful inclusion, despite the absence of guaranteed safety, rights, or institutional recognition.
• Informal groups becoming more visible to an AI system will benefit them rather than make them more governable by states, platforms, funders, or model owners.
• Seven evaluation interviews can support claims about broader network collaboration or systemic impact.
• “Accountable AI” is a property of the assistant itself rather than a consequence of who controls its data, model, and decision channels.

Social Function

Partial truth, transition management, and ideological anesthetic, with a layer of academic prestige signaling.

AMINA may be genuinely useful during the lag period: it can reduce bureaucratic friction and help organizations coordinate under hostile conditions. But its framing anesthetizes the central question—who owns the infrastructure and who becomes economically necessary after cognitive work is automated? It makes a power problem grant-compatible by translating it into design goals.

The Verdict

AMINA is not a counterexample to the Discontinuity Thesis. It is an early artifact of it.

The prototype may improve nonprofit throughput, but it does not restore productive participation, defeat institutional exclusion, or preserve the wage-consumption circuit. It manages the social scaffolding around a population whose formal economic access is already weakening. Its durable value lies in transition intermediation, verification, and coordination—not emancipation.

The marginalized groups receive assistance, not sovereignty. The paper offers a useful prosthesis for a failing institutional body, then mistakes the prosthesis for recovery.