00:00Here's everything we covered in season one of AI at Work in two minutes.
00:06Banking first. AI scores every payment for fraud in two milliseconds, learning what normal looks
00:13like for you and blocking anything that doesn't fit before money moves. American Express improved
00:20fraud detection by 6% across billions of transactions. Bank of America's Erica handled
00:263.2 billion conversations, 98% without a human. Insurance. Lemonade pays simple claims in three
00:36seconds. Progressive paid over 1.2 billion in safe driver discounts through telematics.
00:43AI fraud detection exposed whole criminal rings operating across millions of claims.
00:49Government. The U.S. Treasury used AI to stop $4 billion in fraud in a single year.
00:56Six times the year before. The U.K.'s DWP projects £1.6 billion in savings by 2030.
01:05But the Netherlands' Siri welfare fraud AI flagged thousands of ethnic minority families wrongly.
01:12A court struck it down. The government resigned.
01:16Healthcare.
01:17In 2025, the first fully AI-designed drug completed Phase II A-clinical trials.
01:24AI scribes cut physician burnout from 51% to 39% in 30 days.
01:30The FDA authorized over 1,200 AI medical devices.
01:36Energy.
01:37DeepMind cut Google's cooling energy by 40%.
01:40A gas pipeline reduced turbine failures from six per year to zero.
01:45AI now helps manage the largest grid in North America, serving 67 million people.
01:52And in every single industry, the same catch.
01:55A model trained on biased history faithfully reproduces that bias at scale.
02:02The guardrails, explainability, bias audits, human oversight, the right to appeal are what
02:09make the technology safe.
02:13That's season one, five industries, five game changers, one warning.
Comments