Cornell Professor Says AI Agents Spell Trouble: Bloomberg Crypto

Picture this: an autonomous AI agent launches a memecoin, raises millions, and uses the funds to destabilize a government, says Olga Kharif.

Or offers a bounty — paid in crypto — for hacking an S&P 500 company. Or worse: deploys a smart contract to hire a killer.

This isn’t a Black Mirror episode. It’s a scenario painted by Bill Marino (University of Cambridge) and Ari Juels (Cornell Tech) in their new paper: “Giving AI Agents Access to Cryptocurrency and Smart Contracts Creates New Vectors of AI Harm.”

The warning is stark: the very features that made blockchains popular ...