Good morning. Did you catch the results of the Democratic mayoral primary in New York City yesterday?
Challenger Zohran Mamdani toppled former New York Gov. Andrew Cuomo in the race, besting the better known candidate by about seven percentage points—a big enough gap to prompt Cuomo to concede before voting totals were tabulated.
Why mention a local politics story in a business tech newsletter? Because neither the humans (pollsters, bettors, talking heads) nor the machines (AI chatbots) saw that result coming.
The closest any chatbot got to predicting a Mamdani victory before returns were counted was “a toss-up.” Meanwhile, Polymarket bettors gave Mamdani a 19% chance of winning as recently as Sunday.
It just goes to show that AI is often brilliant at finding the common sense in complexity—and that common sense for certain things isn’t all that common. Today’s tech news below. —Andrew Nusca
Want to send thoughts or suggestions to Fortune Tech? Drop a line here.
Copyrighted books to train AI? Fair. Storing them? Not so much.
A federal judge in San Francisco has ruled that training an AI model on copyrighted works without specific permission to do so was not a violation of copyright law.
U.S. District Judge William Alsup said that AI company Anthropic could assert a “fair use” defense against copyright claims for training its Claude AI models on copyrighted books.
But the judge also ruled that it mattered exactly how those books were obtained.
It was not okay, the judge wrote, for Anthropic to have downloaded millions of pirated titles and maintain a digital library.
The judge ordered a separate trial on Anthropic’s storage of those pirated books. He has not yet ruled whether to grant the case class action status.
With the ruling, Alsup addressed a question that has simmered since before OpenAI’s ChatGPT kick-started the generative AI boom in 2022: Can copyrighted data be used to train generative AI models without the owner’s consent?
Dozens of AI-and-copyright-related lawsuits have been filed over the past three years–most questioning the legal “fair use” of content—to find out.
Alsup’s ruling may set a precedent for these other copyright cases—although it is also likely that many of these rulings will be appealed, meaning it will take years until there is clarity around AI and copyright in the U.S. —Sharon Goldman
U.S. Fed ‘on alert’ for possible Iranian cyberattacks
The U.S. Federal Reserve says it’s preparing for the possibility of Iranian cyberattacks in the wake of the U.S. dropping bombs on nuclear facilities in the country.
“We’re on the alert because we’re a target as well,” Fed chair Jerome Powell said during a congressional hearing Tuesday, adding: “It’s a big issue.”
The chair acknowledged that the central bank had been in communication with peer government agencies with similar concerns. He also said the Fed warned private banks to “be on alert” for cyberthreats.
The concern isn’t unfounded. Iranian hackers were charged last decade with several attempts to hack into banks and other pillars of the American economy.
Last April, the U.S. sanctioned two Iranian companies and several individuals for allegedly working with the cyber unit of the Iranian Islamic Revolutionary Guard Corps to launch phishing and malware attacks on American organizations.
Though the U.S. government doesn’t pinch pennies when it comes to digital security, “You can never, ever be comfortable in this area because the bad guys are always getting better,” Powell said.
So far there have been no known cyberattacks in retaliation for the real use of force over the weekend. Businesses—particularly those who manage or rely on critical Internet infrastructure—remain on alert. —AN
Waymo, Uber arrive in Atlanta
Tired of Tesla robotaxi news? Look no further than Waymo.
The Alphabet-owned autonomous driving outfit began carrying passengers in 65 square miles of Atlanta on Tuesday in partnership with ride-hailing king Uber.
The companies previously partnered on the effort in—where else?—Austin.
Unlike Tesla’s modest robotaxi footprint in the Texan capital, Waymo and Uber have dispatched about 100 driverless vehicles in Austin to date.
In Atlanta, like Austin, Waymo robotaxis are dispatched through the Uber app; riders can opt for a human driver if they’d prefer one.
Waymo’s robotaxi rollout extends beyond its Uber partnership. It has made its own ride-hailing app available in Phoenix, Los Angeles, and a portion of its home metro, the San Francisco Bay Area.
With a quarter of a million paid rides per week, Waymo is considered to be the leader of a still-growing business—in the U.S., anyway—that includes players like Zoox and Lyft partner Mobileye. —AN
More tech
—Job applications via LinkedIn rise 45%. You can thank AI.
—Receive an Apple Wallet alert about F1? Users are “Songs of Innocence”-level irritated about the promotion for the Apple series.
—Mastercard and Chainlink partner. Use your card to buy crypto via fiat-to-crypto conversion.
—Trump’s threatened 25% chip tariff has sparked pushback from a wide range of companies—including automakers.
—Careerbuilder, Monster file for Chapter 11 bankruptcy. The end of an era (and that era is the aughts).
—Microsoft offers Windows 10 security updates for free for an additional year with cloud sync.
—The resurrection of Yahoo. How CEO Jim Lanzone is courting younger users.
—Walmart tests “dark stores,” or smaller warehouses to improve the company’s delivery operation.
Endstop triggered
