1. Warsh Opens AI Review With a Public Trust Deficit and a Productivity Puzzle
Reuters reported that Kevin Warsh, a former Federal Reserve governor, will lead a review of the AI revolution as US public opinion turns more wary. The review will cover productivity, inflation, labor demand, and financial stability. Warsh is not a sitting FOMC voter, but his findings will land inside the Federal Reserve's policy debate. The central bank has cut its benchmark rate by 100 basis points since September 2024. Officials are split on whether AI will lower costs fast enough to justify further easing. Public sessions will include labor groups, small banks, and chipmakers.
The productivity question is the hardest. Major US cloud companies have guided to more than US$200 billion in 2025 capital spending, much of it for AI data centers. Fed staff will try to separate real efficiency gains from spending that merely shifts costs between firms. If AI lowers unit labor costs in services, the disinflation effect could support a 25 basis point cut. If it raises market concentration and pricing power, the Fed may keep rates higher for longer. Warsh has argued that the Fed needs better microdata on firm-level adoption.
Public unease now shapes the policy debate. Reuters reported that the review will include public sessions and consultations with civil society groups. US adults worry about job displacement, data use, and opaque model decisions. The Federal Reserve's Community Advisory Council has flagged small-business anxiety about AI-driven credit scoring. The Bank of Canada has raised similar concerns in its monetary policy report. Warsh will need to explain why a central bank is examining AI at all. His answer will tie the review to the Fed's mandate on prices and employment.
2. Bank Supervision and Tax Administration Face an AI Stress Test
The review will reach beyond macroeconomics into supervision. The Federal Reserve, the Office of the Comptroller of the Currency, and the FDIC are examining AI models used in credit underwriting, fraud detection, and anti-money-laundering systems. In Canada, the Bank of Canada and OSFI are watching the same risks. The IRS and the Canada Revenue Agency are testing AI for audit selection and tax compliance. A model error in credit scoring can deny a loan. A model error in audit selection can trigger a cross-border tax dispute. Warsh will ask whether existing rules can catch those failures.
Payments data are the next pressure point. The Fed operates FedNow and supervises bank holding companies. The Bank of Canada oversees Lynx and retail payment networks. AI fraud tools can cut losses, but they can also create false declines and data concentration. A 25 basis point error in inflation forecasting can swing demand for Treasury duration by tens of billions of dollars. Warsh will examine whether AI vendors create operational resilience risk for banks that rely on a handful of cloud providers. The review will include stress tests and vendor concentration metrics.
Tax administration is part of the same chain. The IRS and CRA exchange information under the US-Canada tax treaty. AI tools can flag transfer-pricing anomalies faster, but they can also produce disputes that land before USMCA panels or domestic tax courts. The CRA has piloted AI-assisted review of small-business filings. The IRS has used AI to identify large partnerships with balance-sheet gaps. Warsh will ask whether central banks should treat tax-model risk as a financial stability issue. That question has no easy answer, and the review will not settle it.
3. Chip Freight, Data Rules and the USMCA Corridor
AI hardware is a physical trade story. Servers and chips enter North America through the Pacific Gateway ports of Vancouver and Prince Rupert, then move on CPKC rail networks toward US Midwest distribution centers. CBP and CBSA screen those shipments at the border. The Ambassador Bridge and the Blue Water Bridge carry auto parts, electronics, and data-center equipment between Ontario and Michigan. A customs delay at one bridge can idle a just-in-time assembly line. Warsh's review will need freight data, not just macro forecasts, to measure AI adoption across the corridor.
USMCA rules will shape the review's trade conclusions. The agreement's digital trade chapter covers cross-border data flows and customs procedures. USMCA dispute panels can hear cases on digital services, customs valuation, and regulatory treatment. CBP's de minimis threshold of US$800 and Canada's lower thresholds affect AI-enabled e-commerce. The USMCA joint review in 2026 will give Warsh a deadline. If AI tools increase small-package volumes, CBP and CBSA will need faster screening. If data localization rules tighten, banks and chipmakers will face higher compliance costs. The review will track both directions.
The Bank of Canada and the Federal Reserve already coordinate on payment systems and financial stability. Warsh will add a trade dimension. AI data centers need power, water, and fiber. Canadian provinces are competing for US$1.2 trillion in global AI infrastructure spending. The Pacific Gateway and CPKC corridor are advantages. So are Canada's clean grid and its access to USMCA markets. The review will ask whether cross-border AI investment is lowering costs or creating new bottlenecks. The answer will depend on CBP and CBSA staffing, rail capacity, and electricity permits.
4. Rate Path Hinges on Whether AI Cuts Costs or Concentrates Them
The monetary policy stakes are direct. If AI lowers unit labor costs in services, the Fed could justify another 50 basis points of cuts over the next year. If AI raises market concentration and pricing power, the Fed may hold rates near 4.25% to 4.50% for longer. Warsh's review will feed into the FOMC's projections and the dot plot. The Bank of Canada faces a similar split. Its policy rate is lower than the Fed's, and Canadian exporters are more exposed to US demand. A divergence in AI productivity estimates could widen the rate gap.
Investors will watch three signals. First, Fed staff data on AI capital spending and depreciation schedules. Second, bank supervision guidance on model risk and vendor concentration. Third, USMCA dispute panels and CBP-CBSA clearance times at the Ambassador Bridge and Blue Water Bridge. A 25 basis point shift in rate expectations can move the US dollar and cross-border equity flows. Canadian pension funds and US asset managers hold large positions in AI-linked infrastructure. The review's findings will shape how they price electricity, rail, and data-center assets.
The review also has to address public trust. Warsh will recommend new data collection, an interagency AI council, and closer coordination with the Bank of Canada and OSFI. He may also propose public reporting on AI use in credit and tax administration. The Federal Reserve cannot regulate AI alone. The IRS, CRA, CBP, and CBSA hold data that the central bank needs. If the review produces only a glossy report, public wariness will grow. If it produces measurable rules and better freight data, the Fed will have a stronger case for its rate path.