Forty Years in the Field, Ten Minutes a Day at the Keyboard
A BBC correspondent who spent decades reporting from Brazil and across Latin America is trading fieldwork for a hammock — but not, it turns out, for a fully unmanned newsroom. Here’s why his instincts are the most valuable asset the site has, and why “AI-driven” can mean something much better than “AI-alone.”
eyesoneurope
There’s something genuinely fitting about this chapter of his career. A journalist who spent forty years learning which sources to trust in Brasília, which official statistics to double-check, and which stories from Georgetown or Paramaribo needed a second phone call before publishing — that instinct doesn’t retire just because the fieldwork does. If AI can now handle the drafting, the research assembly, and the first-pass writing that used to take a newsroom of twenty, what’s left for him to do is exactly the part he’s spent a career getting right: judgment.
That’s not a compromise on the “AI-driven, mostly automated” vision. It might be the version of it that actually works.
Why Investment Content Rewards Exactly His Skill Set
Most content verticals can survive an AI hallucination. A wrong fact in a recipe blog is embarrassing. A wrong fact in an article telling a “potent investor” that a Surinamese mining concession is available, or that a Guyanese port project has government backing, or that Indonesian regulations allow a certain ownership structure — that’s the kind of error that costs someone real money. It’s also exactly the kind of error a seasoned foreign correspondent is trained to catch and a language model, however fluent, is prone to make with total confidence.
Independent research this year on AI chatbots found that over a third of their answers on a comparatively low-stakes topic — election information — contained factual errors, including fabricated candidates and invented scandals. Investment-grade claims about regulatory environments, ownership law, or specific projects in emerging markets are, if anything, harder for a model to get reliably right, since the source material is thinner, less standardized, and changes faster than most training data can track. That’s precisely the terrain where forty years of “check it twice, then check who benefits from you not checking” pays for itself.
The Regulatory Ground Has Shifted — and It Rewards Exactly This Setup
Here’s the part worth knowing before launch: Europe has spent the last two years building legal machinery that treats “substantial human review” as the dividing line between two very different risk categories — and a light, senior editorial pass is enough to land on the right side of it.
Disclosure is now mandatory, but review changes what has to be disclosed. Starting August 2, 2026, the EU AI Act requires that content significantly generated by AI — text, images, audio — be clearly labeled when it hasn’t gone through substantial human review, particularly where the content could influence public debate. Investment analysis on emerging markets sits squarely in that “public interest” category. The good news: content that has gone through a real editorial pass isn’t caught by the same disclosure trigger in the same way. A journalist’s ten minutes of scrutiny per piece is, in a very real sense, doing double duty — protecting readers and satisfying the letter of the law at once.
The liability landscape got simpler, and arguably sharper. The EU’s proposed AI Liability Directive — which would have required proving fault — was withdrawn in 2025 after member states couldn’t agree on it. In its place, the revised Product Liability Directive now treats software and AI systems as “products,” meaning strict liability: if AI-generated content is defective and causes harm, the injured party doesn’t need to prove negligence, only that the product was defective and caused the damage. Member states must write this into national law by December 9, 2026.
Because the company is based in the Netherlands, none of this is hypothetical. This isn’t a question of whether EU rules might reach the site from a distance — the company sits inside the EU, so the AI Act and the Product Liability Directive are the direct, unavoidable legal framework it operates under, regardless of where the hammock is or where the readers log in from.
And ordinary press standards and fraud law apply on top of all that. Every country in question, and most jurisdictions the investors themselves are based in, already has laws against misleading investment promotion that have nothing to do with AI. An AI hallucination doesn’t get a legal pass for being unintentional — Canadian courts already tested that logic in 2024, holding Air Canada liable for its chatbot’s invented refund policy, and European courts are widely expected to follow similar reasoning. A named, credentialed journalist standing behind the site’s editorial judgment is, in this light, not just good practice — it’s a genuine legal and reputational asset.
What “AI-Driven” Can Look Like Done Well
None of this argues against the vision — it argues for building the one thin layer that makes “AI-driven, one person, minimal hours” sustainable rather than fragile:
- Keep a named human editor — him — between the model and “publish,” at least for anything touching a specific law, project, company, or figure. This doesn’t have to mean long hours; it means the discipline of a quick, senior sanity check on the claims that matter most.
- Label AI involvement openly. For a site built on his professional credibility, transparency about method isn’t a liability to hide — it’s the story: decades of correspondent judgment, now paired with AI’s speed.
- Build a visible correction and retraction process before launch. It’s the single best defense against both reputational and legal fallout, and a habit any veteran journalist already knows well.
- Treat any country-specific legal or regulatory claim as requiring a real, named source before publication — “according to [named ministry/law/filing],” not a generated summary standing alone.
- Get local legal counsel in each target country, not just in Brussels. Suriname, Guyana, Brazil, and Indonesia each have their own securities and advertising rules around investment promotion.
- Decide deliberately whether the site is journalism or promotion. A site that curates and analyzes investment opportunities is doing something different — legally and ethically — from one incentivized to attract capital into specific projects. His journalistic background makes the first path the natural and more credible one to build toward.
A Final, Motivating Thought
Here’s the question worth sitting with, hammock and all: if a reader in Amsterdam, São Paulo, or Jakarta discovered that every word on the site had been generated with zero human involvement, would that make the site more interesting to them — or would it quietly evaporate the one thing an investment platform actually sells: the sense that someone accountable stands behind the claims?
In this case, the answer points toward real optimism. “AI-driven” doesn’t have to mean unmanned — it can mean a lifetime of correspondent judgment, applied at the speed AI makes possible, with none of the physical toll of the fieldwork that built that judgment in the first place. Readers may well find that combination more interesting than fully autonomous content, not less: not a machine pretending to have experience, but experience finally freed from having to do all the typing itself. That’s not a limitation on the AI-driven vision. It’s what makes the vision worth trusting.
