Major University study highlights potential $1.2 trillion of wage savings in the US alone
Artificial intelligence is no longer just a talking point at fintech conferences. For the U.S. mortgage industry, it is rapidly becoming a force inside loan origination systems, underwriting desks and servicing shops — and new research suggests it could already perform a startling share of the work those systems contain.
A study from MIT’s Project Iceberg, which simulates the American labor market in granular detail, concludes that today’s AI tools are technically capable of performing tasks worth 11.7 percent of total U.S. wage value — about $1.2 trillion a year. The model represents roughly 151 million workers across 923 occupations, mapped to more than 32,000 distinct skills.
READ MORE: Yes, AI is smarter than you sometimes, but can it help clients with mortgages?
For mortgage professionals, those numbers land close to home. Few sectors blend financial analysis, document review and administrative routines as tightly as home lending does.
A microscope on the work behind every loan
Project Iceberg does not predict next year’s unemployment rate. It does something subtler: it asks, for each occupation in the economy, what fraction of its skills are ones that AI has already demonstrated it can perform in at least one real context.
To do this, the researchers treat each worker as an “agent” with a bundle of tasks and skills, then match those against thousands of existing AI tools. The resulting measure, the Iceberg Index, captures what the authors call technical exposure — “where AI can perform occupational tasks, not displacement outcomes or adoption timelines.”
In practice, that means the 11.7 percent figure is not a forecast of layoffs. It is a map of where companies already have a choice: continue to use people, deploy machines, or design some hybrid of the two.
In the mortgage business, where an application file can pass through dozens of small, rule‑driven steps, that choice is quickly becoming unavoidable.
The real action is below the surface
It would be easy to assume that the brunt of AI’s impact falls on software engineers and data scientists. Project Iceberg offers a different picture.
When the researchers look only at current AI adoption in computing and technology occupations, they find exposure of about 2.2 percent of total wage value — roughly $211 billion. They call this the “Surface Index,” and describe it as “only the tip of the iceberg.”
Beneath that visible tip lies a much larger mass of potential automation. Technical capability “extends far below the surface through cognitive automation spanning administrative, financial, and professional services,” they write, bringing total exposure to 11.7 percent and about $1.2 trillion in wages.
One summary of the research notes that high Iceberg scores in several states are driven by “cognitive work—financial analysis, administrative coordination, and professional services.” That phrase could just as easily describe the back office of a major mortgage lender.
The mortgage file, as AI already sees it
The Iceberg report does not list “loan officer” or “mortgage underwriter” among its villains. Instead, it focuses on the kinds of tasks AI has mastered: reading documents, extracting data, applying rules and generating standard text.
Those are not abstractions in this industry.
Document and data processing
The researchers point out that financial institutions now deploy AI for “document processing and analytical support,” while healthcare systems automate “administrative tasks.”
Mortgage operations combine both patterns. Every day, systems — and the people behind them — read pay stubs, tax returns, bank statements, credit reports, purchase contracts and appraisals. They extract income, debts, assets and red flags into structured fields. They classify documents and chase missing pieces.
These are precisely the tasks Iceberg tags as highly exposed.
Rules, thresholds and triage
The same is true of policy checks and triage. Eligibility against product guidelines, calculations of debt‑to‑income ratios and stress tests, and routing of clean files to one queue and messy ones to another are all rules‑based operations. They are technically demanding to implement, but conceptually straightforward for software that thrives on patterns.
Iceberg’s message is that current AI can already handle a surprising share of such work.
Scripts, not stories
Then there is the writing. Much of the industry’s written output — pre‑approval letters, disclosures, adverse‑action notices, conditions lists — follows templates set by compliance and legal teams. The variables change; the language does not.
To a generative model, this sort of constrained writing is low‑hanging fruit.
None of this means that loan officers, processors and underwriters suddenly vanish. It does mean that the slices of their day devoted to data entry, document classification and boilerplate communication are under real pressure.
A profession at a crossroads
If Iceberg was designed for policy makers, its questions now belong on the desks of mortgage executives.
One aim of the project is to help governments “identify exposure hotspots, prioritize training and infrastructure investments, and test interventions before committing billions to implementation.” That same logic applies to lenders and brokerages contemplating new technology budgets.
Three questions loom.
Where is your own iceberg?
Because Iceberg is built on skills rather than job titles, exposure varies inside roles. A seasoned underwriter who spends half her time negotiating edge‑case files with originators is differently exposed from a colleague who spends most of the day verifying standard documents against checklists.
For any given shop, the first task is to deconstruct jobs into tasks: data entry, document review, simple versus complex decisions, direct conversations with borrowers and real‑estate agents. Only then can leaders estimate what share of their wage bill sits in work that looks like the document processing and administrative tasks the report highlights.
What do you want to do with 11.7 percent?
The Iceberg number invites a blunt strategic choice. Some institutions will use AI primarily as a cost‑cutting tool, shrinking operational headcount as systems take over routine work. Others may hold headcount steadier, allowing AI to absorb volume growth while redeploying people into areas that resist automation: complex structuring, retention, outreach to underserved communities, and the messy human business of helping borrowers through financial stress.
For a profession that remembers the aftermath of 2008, the direction of travel will matter as much as the destination.
Can you explain your algorithms — to regulators and borrowers alike?
The researchers note that traditional indicators such as GDP, income and unemployment explain “less than 5% of this skills-based variation,” underscoring why new indices are needed to capture AI exposure. The regulatory questions, however, are old ones: fairness, transparency, accountability.
An AI‑heavy operation must still answer familiar queries from examiners and consumers: Why was this loan declined? Why did that one sail through? How are protected groups treated? Iceberg’s emphasis on pre‑emptive planning is a reminder that these questions should be addressed before automated decisions drive outcomes at scale.
A shrinking window
Project Iceberg is, in some sense, a warning system. It exists to help societies prepare for change before it appears in unemployment statistics and shuttered offices. The researchers put it starkly: “The window to treat AI as a distant future issue is closing.”
For the American mortgage profession, that window may feel particularly narrow. The industry sits at the intersection of finance and paperwork, regulated with increasing intensity and competing on speed as much as on rate.
AI is already capable of taking on a meaningful slice of the work behind every closed loan: the scanning and sorting of documents, the first pass at eligibility, the assembly of disclosures. The open question is what the industry chooses to do with that capability — whether it becomes a way to move more quickly and humanely through the home‑buying process, or simply a quieter, colder form of efficiency.
As Project Iceberg suggests, the technology is no longer the bottleneck. The decisions are.
