Hi. I'm Gabriel Elizondo. For 21 years I've been a journalist with Al Jazeera English, in Washington DC, South America and New York, and today I cover diplomacy and global affairs from the United Nations. I've reported from more than a dozen countries. I built a bureau in Brazil and traveled to all 26 of its states. I've reported on many different stories, including elections across South America, the war in Ukraine, the pandemic in New York, and illegal gold mines and drug labs deep in the Amazon in Brazil and Peru. I was in Washington for the launch of Al Jazeera English's coverage of the Americas.
I'm curious about other cultures and countries, and I've been lucky enough to visit 64 of them so far. Travel always rejuvenates me.
Journalism? It was by accident. I studied global conflict resolution in undergrad and international economics in grad school. I thought I would be a diplomat or an international banker. But I got pulled into reporting along the way and never pulled myself out. No regrets.
Nothing will ever replace on-the-ground, exclusive reporting. I've spent two decades seeing that firsthand.
As a correspondent, the most intimidating and exhilarating moment was walking off a plane in a foreign country, sometimes without speaking the language, and having to find a story. Often on a tight deadline. Sometimes I had to file only an hour after touching down.
When I started building with AI this year, it gave me that same feeling of intimidation and exhilaration.
I've now built tools I use on my beat, and that other people use too: an app that tracks hundreds of speakers, speeches and events during the UN General Assembly, and an agent I designed that has been following the race for the next UN Secretary-General on its own since August. My Building page has the full list. I build with Claude Code, Codex and Grok Bot, and I write about what works and what doesn't.
The biggest thing I've learned is that when AI is turned on, the editor's job doesn't turn off. It just moves. The journalist sets the standards, checks the decisions, catches the edge cases and changes the rules when they fail. Twenty years of deadlines and fact-checking turned out to be the right training for it.
The instinct I bring is simple: make things that are honest about what they can and can't do. Systems that flag uncertainty instead of hiding it, hold up under pressure, and earn trust rather than assume it.