Systemic data governance gaps obstruct equitable agentic AI in finance: structural, regulatory, and cultural barriers revealed
Original framing: “Data readiness for agentic AI in financial services” — MIT Technology Review
Agentic AI in finance relies on flawed assumptions about data quality, often treating historical biases as 'noise' rather than structural features of financial systems. Studies show that algorithmic credit scoring replicates discrimination even when trained on 'clean' datasets, revealing that the problem is not data readiness but systemic inequity. The scientific literature on fairness in AI highlights the need for causal models that account for historical power dynamics, not just statistical corrections.
The 'data readiness' narrative in financial AI is a Trojan horse for neoliberal extraction, disguising centuries-old power structures as technical necessities.
From colonial banking to redlining, financial systems have always been about control—now repackaged as 'agentic' efficiency. Indigenous data sovereignty, African relational economics, and Buddhist critiques of accumulation all reveal that the problem is not data scarcity but power asymmetry. The solution lies not in better algorithms but in dismantling the infrastructures of exclusion, replacing them with systems of accountability and reciprocity. Hermes, the trickster, reminds us that the most 'sophisticated' AI is still a human construct—and thus subject to human fallibility, humor, and justice. The future of financial AI must be co-designed with those it has historically harmed, not for them.