Tariffs and AI: The Blind Spot Nobody’s Pricing Into Your Portfolio

I’ve spent more than 15 years in front of trading screens and treasury desks, and this week I had a strange feeling: I was reading two headlines that belong on the same front page, and nobody seems to be putting them together. On one side, the US bumped Colombia’s tariff to 12.5% starting July 24 — flowers, apparel, chocolate, cosmetics, palm oil, all suddenly more expensive to sell into the world’s biggest market. On the other, 91% of global strategists are betting that artificial intelligence will be the single factor that defines market performance for the rest of the year. Two stories, one reader: whoever has to decide what to do with their money this week.

What’s Actually Happening

The tariff isn’t a technical footnote. It hits three out of every ten dollars Colombia exports to the US, and it already cost exporters roughly $80 million extra in July alone. The official reason — that the country lacks “sufficiently effective” controls against goods made with forced labor from third countries — reads almost like a footnote next to the real effect: Colombia now sits 2.5 points above Ecuador, Mexico, and Guatemala in the same market. Big funds noticed immediately. Bank of America’s August fund manager survey shows preference for Colombian assets dropped from 37% in July to 30% this month, and only 16% of managers plan to raise their equity exposure over the next six months — the lowest reading since October 2024.

Meanwhile, Natixis Investment Managers’ early-August survey found that 79% of strategists expect AI-linked volatility to spill into sectors that aren’t even tech, and 45% expect a return on that AI investment within just twelve months. In other words: the same crowd trimming exposure to Colombia over political and trade risk is, at the same time, doubling down on a model — generative AI — that hasn’t yet proven it can read political risk at all.

What Nobody’s Telling You

Here’s the blind spot: the AI models feeding today’s screeners, research summaries, and even portfolio recommendations are remarkably good at predicting price patterns, but they don’t smell a tariff coming before it’s signed — and that gap is precisely the seed ViSFI and the Método CRITERIO grew from. No training dataset captured that the USTR would use a forced-labor clause as trade leverage against Colombia. That kind of risk — the kind that’s born in a Washington office, not a spreadsheet — is still exclusively human territory. And yet 59% of advisors in the region already use AI routinely for research, communication, and portfolio analysis, even as 71% admit, when asked directly, that handing full investment counsel to a machine is an unnecessary risk. They use it, but they don’t fully trust it — and they’re letting it get closer to the wheel anyway.

Where AI Earns Its Keep (and Where It Doesn’t)

I said it before in my piece on AI in the age of financial analysis: the tool is never the problem — the problem is trusting it without understanding it. After watching enough tariff cycles and tech fads come and go, my read is simple: AI wins when the problem is volume — cross-checking a hundred quarterly reports, catching a cash-flow anomaly, summarizing thirty pages of a prospectus. It loses, badly, when the problem is context — understanding why a government decides to squeeze one specific sector, or what happens to an investor’s confidence when forced-labor headlines start running next to their country’s name. That’s exactly the moment where building your own financial judgment — not outsourcing it — is worth more than any model. If you’re using AI tools for this week’s decisions, use them for what they’re actually good at, and keep the call on political, regulatory, and reputational risk for yourself.

Are you asking an algorithm to understand something even the USTR didn’t fully explain? That’s worth sitting with before your next portfolio decision — and it’s exactly the kind of exercise we run daily in ViSFI’s Daily Briefing, where we connect these signals before they become headlines.

If this was useful, share it — and if you want to go deeper on how we’re reading the intersection of trade policy and AI in the markets, ViSFI’s Daily Briefing tracks this signal every day.