Ex-Harvard Professor Builds An AI Tool That Can Predict & Stop Wars — Eh, Sure Or Not?

Meet North Star, a predictive AI system that simulates world leaders' decisions.

Enlarge text
Logo

As geopolitical tensions rise, one former Harvard academic reckons he's built an AI tool that could help avoid the unthinkable

Meet North Star, a predictive AI system that simulates world leaders' decisions to foresee what could happen in escalating conflicts.

Its creator, Arvid Bell — an ex-Harvard professor and co-founder of political risk firm Anadyr Horizon — introduced the tool recently at the AI+ Expo in Washington, the US.

"I want to simulate what breaks the world. I don't want to break the world," he said, according to Interesting Engineering.

Image via New Straits Times / ROBERTO SCHMIDT / AFP

North Star creates digital twins of political leaders, then drops them into simulated scenarios like military strikes or economic sanctions

The AI plays out the consequences, kind of like a high-stakes version of The Sims, but with nuclear implications.

Back in 2022, the software predicted a 60% chance of Russian escalation following a no-fly zone over Ukraine.

It even generated a hypothetical intelligence brief echoing Russia's wartime rhetoric.

Image via Kevin Ku / Pexels

While North Star claims to help avoid war, it also happens to be big business

Anadyr sells the software to corporate risk teams and has already drawn investor comparisons to the early days of climate tech.

Business Insider reported recently that the "peace tech" industry could be worth billions, piquing the interest of investors with big wallets.

Whether this tech actually promotes peace or just profits, the idea of governments relying on AI to make war decisions is… unsettling. Because when machine logic meets messy human geopolitics, what could possibly go wrong?

Follow SAYS Tech on FacebookInstagram, & TikTok for the latest in tech in Malaysia and the world!

Read more #tech stories:

Read more trending stories on SAYS

You may be interested in: