This researcher says the incentives for developing AI are all wrong

PLUS: The latest AI and tech news.

View this email in your browser  |   Manage newsletter subscriptions By Jennifer Conrad | 11.11.21 Was this email forwarded to you? Subscribe here. Happy Veterans Day! I've spent the past two days glued to my computer screen for RE:WIRED, a series of conversations about humanity's biggest bets in health care, art, artificial intelligence, and cybersecurity.     Timnit Gebru Says Artificial Intelligence Needs to Slow Down All too often, technologists don't get to build artificial intelligence models the right way, based on a careful weighing of the pros and cons, AI researcher Timnit Gebru told Tom Simonite during a RE:WIRED conversation on Tuesday. In the rush to publish new research and push out ever larger models, "we haven't had the time to think about how it should even be built because we're always just putting out fires," she said. As Simonite recounted in a WIRED cover story in June, until a year ago, Gebru was a researcher at Google and co-led a team dedicated to ethics in AI. She says she was forced out in a dispute over a research paper that detailed the ways things can go awry with large language models, which power services like machine translation and Google search. The experience left her convinced that the incentives behind current AI research are all wrong, meant to "help the Defense Department figure out how to kill more people more efficiently" or make more money for multinational corporations. Gebru plans to open an independent interdisciplinary institute for AI ethics and accountability on December 2—one year to the day from her ouster from Google. "The hope is that instead of just constantly critiquing technology after the fact … we can also maybe model a positive example for how we should do AI research," she said. Watch the discussion with Timnit Gebru.     Deep(er) Learning: RE:WIRED Edition Kai-Fu Lee and Yoky Matsuoka Imagine AI's Potential for Good Sinovation Ventures CEO Kai-Fu Lee says the "black-box problem"—the idea that it's impossible to understand how an AI makes its decisions—can be overcome. Rather than expecting to perfectly grasp the AI's complex mathematical logic, we should seek a "reasonably accurate" explanation a human can comprehend. Jen Easterly Wants Hackers to Help US Cyber Defense "All it takes … is a little bit of money" to launch a ransomware attack, warns Jen Easterly, director of the US Cybersecurity and Infrastructure Agency. But she's optimistic the government can work with the private sector—and the hacker community—on cybersecurity. Who Owns the Vaccine? A Conversation With Moderna's Stéphane Bancel and Dr. Nahid Bhadelia mRNA vaccines against Covid-19 were created in record time. The next steps are making them affordable and available to everyone—and developing mRNA treatments for other illnesses, including cancer.     Need to Know The Infrastructure Bill: 5 Key Takeaways The $1.2 trillion plan includes money to expand broadband access, improve walking and cycling trails, and upgrade public transportation. The Turing Test Is Bad For Business AI should be harnessed to enhance productivity and complement humans, rather than trying to imitate them. 'AR Is Where the Real Metaverse Is Going to Happen' "I don't need to make a conference room look like a cartoon Tahiti." John Hanke, CEO of the company behind Pokémon Go, says developers should be focused on enhancing everyday life, rather than creating full-on 3D worlds. The Hidden Dangers of 'Buy Now, Pay Later' Apps Services that let people pay in installments for everything from rent to Gucci bags are soaring in popularity and valuation. But they also make it easy for users to buy more than they can afford.       So, This Happened Streaming platform Netflix faces more defamation lawsuits than major news outlets—and the suits raise questions about whether it should be viewed as a publisher, a distributor, or something else entirely. (The Hollywood Reporter) Displaced people in the Democratic Republic of Congo are using bitcoin to create a grassroots economy. (TechCrunch) Crop-monitoring drones and horses equipped with poop sensors are giving Mount Vernon, George Washington's old farm, a high-tech update. (Fast Company) Is a global energy crisis in the forecast this winter? (The Cipher Brief)     Until Next Time   A cartoon by Lars Kenseth, with Collect Cartoons. GET WIRED A helicopter on Mars. What to expect post-Covid. The next electrifying/terrifying leap in A.I. explained. Get 1 year of unlimited access to WIRED for less than $1 per month. That includes Steven Levy's exclusive newsletter and a free tote! Subscribe now.             This e-mail was sent to you by WIRED. To ensure delivery to your inbox (not bulk or junk folders), please add our e-mail address, wired@newsletters.wired.com, to your address book.   View our Privacy Policy Unsubscribe or manage your newsletter subscriptions   Copyright © Condé Nast 2021. One World Trade Center, New York, NY 10007. All rights reserved.