On frontiers, tsunamis, and writing about AI from inside the wave.

My New Book Arrives September 3: Foreign Policy at the AI Frontier - How AI is Reshaping Progress, Power and Peace

---

In technology today, everyone is talking about the frontier. Frontier models. Frontier labs. Frontier AI. Frontier risk. The word marks the edge of what machines can do — the terrain where capability often outpaces regulation, and ambition often outpaces understanding. But there is another kind of frontier, one much older than any algorithm. Foreign policy has always been conducted at frontiers: at borders and beyond them, where sovereignty meets sovereignty, where diplomacy is tested and the fate of nations is decided.

My new book, Foreign Policy at the AI Frontier: How AI is Reshaping Progress, Power, and Peace (published on September 3, 2026 by Routledge), sits precisely where these two frontiers meet — tracing the intersections between artificial intelligence (AI) and foreign policy (FP), and examining how AI is simultaneously a subject of diplomatic negotiation, a tool for conducting international relations, and a force reshaping the global order.

The book is shaped by a decade of work and projects with foreign ministries, embassies, consulates, and partners across industry, civil society, and academia — and by a conviction I have come to hold more firmly with each passing year: that AI can and must serve to sustain and strengthen peace, not to wage war.

That's the story I want to tell here. Not the book itself —I hope you will read it for yourself — but the story of how it came into being, why it took the shape it did, and why, in a very real sense, the conversation it opens can never truly be finished.

The Tsunami and A Huge Wave You Have to Learn to Swim In

The book did not begin with a single moment or a single metaphor. It began as a slow accumulation — of themes, conversations, encounters, and above all questions — gathered over more than a decade of work at the intersection of many frontiers: technological and diplomatic. More specifically it grew from the frontier where AI meets peace, humanitarian action, international development and foreign policy. The same questions on AI opportunities and risks kept surfacing in different rooms, in different languages, on different continents.

One moment stands out, though, because it gave a name to what we have been sensing for years. While teaching at the Diplomatic Academy of Chile, a young diplomat-in-training described AI to me as a tsunami — a huge wave you have to learn how to swim in. Colleagues in Washington used the same word and similar metaphors. Then Nairobi. Then Bangkok. Then Brasília. Diplomats on four continents, none in conversation with each other, all reaching for the same image.

That was when the shape of this book finally came into focus. If seasoned practitioners of statecraft — people trained to speak carefully, in measured terms — were all reaching for the language of natural disaster, something was already happening beneath the surface of world affairs.

But every tsunami also reveals a new frontier — uncharted territory where the rules of statecraft have yet to be written, and where diplomats are already learning to swim. This book is an invitation to do two things together: to learn to swim in the wave that is already upon us, and to shape — collectively, deliberately — the frontier that will emerge once the water pulls back.

What's Inside the Book — A Few Teasers

Without giving away the arguments themselves, here is a glimpse of the terrain the book covers, organized around its three themes — Progress, Power, and Peace.

It asks what it means when code replaces cabinet officials, and what precedents that sets — and breaks. It explores the rise of AI candidates, AI parties, and the early experiments in automated democracy happening in unexpected corners of the world. It traces the arc from Sputnik to DeepSeek, examining how technology shocks reset global power and why the current one may be different from any that came before.

The book takes readers into the dangerous convergence of nuclear weapons systems and AI decision-making — a convergence that pushed the Doomsday Clock to 89 seconds to midnight, the closest to catastrophe it has stood in its 78-year history.

It examines how satellites are becoming digital witnesses to war crimes, transforming human rights documentation and the very meaning of evidence. It explores the surprising, underreported ways algorithms are already shaping peace mediation, peacekeeping, and the Women, Peace and Security agenda.

Running through all of this is a chapter on the ethical and governance questions that no diplomat can any longer avoid. What happens when the models informing a foreign minister's briefing are black boxes whose reasoning cannot be audited? When AI systems hallucinate facts into diplomatic cables, or quietly rewrite history through the biases embedded in their training data? And how does a diplomatic system built on deliberation govern a technology moving at the speed of silicon — where a single model release can reshape the strategic landscape faster than a treaty can be drafted?

Each one of these topics has a full chapter behind it. And each one, in different ways, sits at that intersection I keep returning to — where the technological frontier and the diplomatic frontier meet.

The Moving Map: Writing on AI as AI Kept Racing Ahead

Here is what makes this book strange to have written: many of the phenomena it describes did not exist when I began writing. When I started, "tech diplomat" was a job title held by a handful of pioneers; now dozens of countries are appointing them, and I argue in the book that soon every diplomat will need to be one. "Agentic diplomacy" — the idea that AI agents themselves might enter the negotiating room — sounded speculative when I first wrote about it; by the time I was editing, it was already being tested. ChatGPT and DeepSeek were not yet household names. "Sovereign AI" was a niche debate, now it’s a dominating topic across the world of AI. Several chapters I had to rewrite two, three, sometimes four times.

I have written a book before, before so called “generative AI revolution” (AI for Peace). That one gave me the gift most authors take for granted: time. Months and months of reflection on events, projects, and lessons already embedded in a stable, pre–generative-AI world. I could revisit ideas, let them settle, watch them ripen. This book was different.

Almost every morning of the writing process unfolded the same way: coffee, the headlines, and the quiet suspicion that the task was impossible. Overnight, a new frontier model had been released. A lab had crossed a benchmark that only weeks earlier had seemed out of reach. A company had announced a capability that turned a hypothetical scenario in my draft into a real one. A new player — sometimes from an unexpected corner of the world — had entered the race.

And it was not only innovation that kept moving. Governance was moving too, sometimes just as fast: new export controls, new AI acts, new declarations, new institutions being invented in real time. A summit had produced — or failed to produce — an AI framework. A paragraph written on Tuesday could be out of date by Thursday.

This change ran in both directions. For every AI office being launched, another was being quietly closed. For every task force convened, another was being disbanded. In this field, the speed of creation is matched by the speed of dissolution — and the institutional map of AI diplomacy is being drawn and erased at almost the same pace.

There were moments — many moments — when I wondered whether writing a book about AI was itself a category error. In the end, I made peace with a simple truth: this book is a screenshot in time. It is the most complete map I could draw of a landscape still forming under our feet. It is not the final word on AI and foreign policy. It cannot be. No book can be.

Why This Is the Beginning, Not the End

That is precisely why the book cannot stand alone. Every week now, something new happens in this space that deserves attention — a new AI actor entering diplomacy, a new governance debate, a new use case in peacebuilding or crisis response, a new escalation in the AI arms race.

This time, I am also dedicating far more effort to shaping the conversation around the book than I did with my previous one. That is why I built this website — “Foreign Policy AI” — a companion to Foreign Policy at the AI Frontier, and a place to keep the conversation moving after the book goes to print. Here you will find book content, endorsements from colleagues and reviewers, more about me and the path that led to this book, and the growing calendar of talks, book launches, and public conversations already scheduled across September, October, and November, with more being added each week.

If any of this resonates — if you sense, as I do, that the choices being made now in AI and foreign policy will echo for generations — then I would be honored to have you not only as a reader, but as someone to think, question, and shape this conversation with. Subscribe to the website to receive updates on new writing, upcoming events, and where the book conversation is heading next; follow me on LinkedIn for the regular Is AI Coming in Peace? newsletter; and read the book — because the future we are about to inherit deserves more than our attention; it deserves our engagement.

Because in the end, we will not get the future we fear. We will not get the future we hope for either. We will get the future we prepare for — the one we are willing to shape, deliberately, together, before someone else shapes it for us.

This book is my contribution to that preparation. I hope you will join!

---

Branka Panic, Author, Foreign Policy at the AI Frontier - How AI is Reshaping Progress, Power and Peace; Founder and Executive Director, AI for Peace