The Rise of AI | No Reservations | 2

0:00Before we get started, a brief content warning. This episode contains a discussion of gun violence. Please be advised.

0:16It's late October 2015 in Cambridge, Massachusetts. Joy Boulamwini slams her hand against her desk in frustration. The intricate Lego display spread across the table shakes from the impact. Boulamwini takes a deep breath. She can make this work. Boulamwini is in her first semester of a graduate program at MIT. She's working on an assignment to build a piece of technology inspired by science fiction. Boulamwini is making what she's dubbed an aspire mirror. The idea is that when a user looks in the mirror, it will reflect back an image superimposed on the user's face that will inspire them. It could be the face of a lion, or tennis superstar Serena Williams, or even a quote or a symbol.

1:07Bulemwini is using facial recognition software to track the user's face to know where to place the inspirational image. But try as she might, she can't get the software to work. Over and over, a message appears that reads, No face detected, even though she's staring right into the mirror. Bolamwini adjusts her chair and angles her chin down, hoping the change in position will help the software recognize her. No face detected blares in red lights across the top of the mirror. Bolamwini sighs. Why won't you work? She spins around in her chair. She doesn't know what she's going to do. The assignment is due soon, and without reliable facial recognition software, it won't work.

1:54As her chair comes to a stop, Bolamwini spies a white Halloween mask sticking out of her bag. She'd brought it to a friend's party. Bolamwini grabs the mask and slips it over her face. She turns and looks into the mirror. Staring back at her is a lion, regal and proud. Behind the mask, Bolamwini wrinkles her brow. She moves her head to the left and then the right. The software tracks the white mask perfectly. She shoves the mask on top of her head and looks back into the mirror. No face detected, the computer blares. Bulemwini slips the mask back down. A lion stares back at her. Bulemwini closes her eyes.

2:39She knows exactly what's happening. In order to learn to recognize what's a face versus what's not, artificial intelligences are fed massive data sets. But if there's a flaw in the data set, that flaw will come through in its performance. So if there were only a few black faces in the data set that trained the algorithm, it would explain why the mirror didn't pick up Bula Mwini's face. It literally hadn't been taught to recognize her as a person. Bula Mwini's stomach sinks. She started coding because it was an activity that felt divorced from the real world. But it's clear that AI is full of human prejudice.

3:23Bulemwini begins taking notes, writing quickly as questions and thoughts flood her mind. She's determined to study the bias she just experienced. This technology is being sold as objective and fair and already being deployed to law enforcement. There will be devastating consequences if facial recognition AI cannot reliably identify darker-skinned people. Boulamwini's not the only scientist seeing problems with what's being touted as the next big thing. All the major AI companies will soon find themselves in difficult positions. And how they respond will determine who takes the lead.

4:12When planning for your future, you want someone with a history of keeping their word year after year. For nearly 160 years, Pacific Life has been a trusted name in the industry. But that isn't just a number. It's experience that matters. It's 160 years of promises held, helping generations retire with confidence, protect their loved ones, and plan for whatever comes next. Whether you're looking for life insurance, employee benefits, or retirement income solutions, when your future is on the line, you want history on your side. And believe me, Pacific Life has been there, always there through changing times, always focused on your needs.

4:50Ready to secure your tomorrow? Ask a financial professional how Pacific Life can help you feel prepared for what's next. Pacific Life Insurance Company, Omaha, Nebraska, and in New York, Pacific Life & Annuity, Phoenix, Arizona. Because with Pacific Life, you're not just planning for the future, you're partnering with trusted experience. Closing the books, getting your people paid, and bringing on new hires. Running a small or mid-sized business can be exciting and also a little chaotic. Workday Go makes simplifying your business, well, simple. Imagine all the important aspects of your company, HR, finance, and payroll, all on the AI platform.

5:32No more juggling multiple systems, no more worrying about growing too fast. Just the full power of Workday Go, helping small to mid-sized businesses like yours scale and run more smoothly. Think about what that means. Seamless onboarding for new team members, real-time insights at your fingertips, and payroll that works perfectly every single time so you can focus on the big picture and go after your big ambitions. And with Workday, you can activate quickly in as little as 30 to 60 business days. So simplify your business. Go for growth. Go with Workday Go. From Wondery, I'm David Brown, and this is Business Wars.

6:43On our last episode, the major tech companies, including Google, Microsoft, and Facebook, all made significant investments in artificial intelligence. They concluded that AI would change the world in the same way the Internet and the smartphone once had. Google quickly became the clear leader. Part of their advantage with AI companies and researchers came from their reputation as a tech giant with ethics. Scientists were concerned about the impact of AI on society and wanted to work with a company that took those concerns seriously. One that didn't just push for constant advances, no matter the consequences.

7:20However, a new player came on the scene, funded by Silicon Valley celebrities like Elon Musk and Peter Thiel. Called OpenAI, it was structured as a non-profit. The company declared it would put a strong emphasis on ensuring any AI they develop would be ethical and beneficial to humanity. But as AI continues to develop at a breathtaking rate, more and more people are starting to have doubts that the ethical guardrails any of these companies had put in place were enough. On this episode, each of the major companies comes up against the dark side of AI. They have to decide how to weigh their ethical concerns with their desire to get ahead.

8:05This is Episode 2, No Reservations.

8:18It's September 2017 in Palo Alto, California. Fei-Fei Li logs into a video conferencing software from her office at Stanford University. Li is the director of Stanford's Artificial Intelligence Lab, as well as the lead scientist for Google's cloud computing division. And today, some Google colleagues have asked her to join a meeting about an exciting but controversial new opportunity for the company's AI division. One by one, her colleagues log on, including the head of Google's Defense and Intelligence sales department. Thanks for coming, everyone. Fei-Fei, thanks for making the time to meet with us.

8:55I know we're all busy, so let's just dive in. Sounds good to me. As we're all aware, Google's about to sign a contract with the Department of Defense for an AI initiative called Project Maven. The idea is that the military will use AI to analyze satellite images as well as video footage shot by drones. Now, the question is whether we should publicize this new deal. To get trusted with a DOD AI contract is a big deal, and it could help us win other bigger contracts. But having said that, people are uneasy about military use of AI, so I guess we'll get some pushback. The head of sales glances at Lee's box on his screen.

9:34Feifei, as someone who's really plugged into the AI community, I'd love to get your thoughts. Well, first, I want to say that this is an amazing effort to secure the contract, so congratulations to your whole team. That said, I do think we should exercise extreme caution when promoting it. I think we could do a PR push about the DoD working with our cloud computing services, but I think we should avoid any mention of artificial intelligence. It's just too much of a hot-button issue right now. Several staffers on the call nod in agreement, but one frowns. Of course, I respect Fei-Fei's opinion immensely, but I just want to throw out a different perspective.

10:11It seems to me that it's highly unlikely we'll be able to keep Google's involvement with Project Maven under wraps forever. At some point, you know it's going to leak. So maybe it's better that we just put the story out there and frame it so we're not forced to play defense. The head of sales nods. That's a good point. And how would you frame it? Well, when General Mathis met with us about this collaboration, he said one of the reasons he was talking to us was because of our commitment to corporate responsibility. So I think we could also talk about how the fact of the matter is that bad actors are going to use this technology, so it's vital that the good actors stay one step ahead.

10:51That's what we're helping to do, right? Lee raises her finger, indicating that she wants to speak. I understand that point, and maybe a month ago I would have agreed. But just a few weeks ago, a petition was circulated that called for the UN to ban AI weapons. And several of our company's most prominent researchers, including Jeff Hinton and Demis Osavis, have been outspoken about their distaste for military AI. I just think this contract will be like throwing red meat to the media. They'll paint us as hypocrites, betraying our ethical standards. The head of sales sighs. Well, I don't think we're going to make a decision about this today.

11:28We'll keep talking about it over the next several days. Ultimately, Google decides not to publicize that it's working on Project Maven. The company doesn't even announce internally that they're working with the DOD. But in early 2018, word does get out. And the reaction is not good. Employees learn that Google is assisting the Pentagon with deploying drones. And some are furious. Protests spread throughout the company. The team of engineers assigned to the project refuse to work on it, risking their jobs. Upper management holds an emergency town hall and tries to assure employees that the company is not involved with drone strikes and is not doing anything that could be considered an offensive military attack.

12:15But the turmoil within the company continues to grow, and media outlets pile on. Some Google employees circulate an internal petition demanding Google cancel Project Maven. More than 3 ,000 employees sign it, including some of Google's most prominent researchers and one of the founders of DeepMind. Google's reputation as the ethical tech company is in danger. And they're not the only company facing pushback. Facebook founder Mark Zuckerberg is called to testify in front of the Senate. Ostensibly, the hearing is about how the U.K. company Cambridge Analytica used private Facebook data to target voters on behalf of the Donald Trump campaign.

13:04But the questions soon become more broad-ranging, with numerous senators asking Zuckerberg about the company's use of AI and the ethical guidelines they have in place. Is your company, you mentioned principles, Is your company developing a set of principles that are going to guide that development? And would you provide details to us as to what those principles are and how they will help deal with this issue? Yes, Senator. I can make sure that our team follows up and gets you the information on that. And we have a whole AI ethics team that is working on developing basically the technology. It's not just about philosophical principles.

13:37It's also a technological foundation for making sure that this goes in the direction that we want. Thank you. Zuckerberg remains bullish on AI, convinced that within 5 to 10 years, it will be able to detect hate speech, a nuanced linguistic endeavor, and bragging that 90 % of Al-Qaeda and ISIS content is flagged by AI. But the line of questioning makes clear that Facebook is losing the trust of politicians, and by extension, the American public. The company won't be granted much leeway if there are any mishaps. with its AI.

14:20It's early 2018 in San Francisco. Sam Altman peers over at his colleague Greg Brockman, relieved that he looks as skeptical as Altman feels. In front of him, tech billionaire Elon Musk gestures emphatically as he rants. It's embarrassing. Google's making breakthrough after breakthrough, and what are we doing? Sitting around on our thumbs. Altman largely tunes him out. Three years ago, Altman, Brockman, Musk, and a slew of other tech visionaries founded OpenAI, a non-profit dedicated to developing ethical AI that is beneficial to humanity. Musk is on the board and is one of their biggest donors.

15:00He's pledged a billion dollars. But today, he's not happy. Musk uncrosses his leg, setting his foot down with a thunk. I think it's time to admit that this experiment isn't working. We are falling fatally behind Google, and if things don't change, open AI will never catch up. I don't think that's entirely fair. We're making progress on our robot that can solve a Rubik's Cube, as well as an AI that can beat some pretty complex video games. And we're making big strides in text generation. Progress, but not leaps. We said we were going to be the American version of DeepMind, but we haven't done anything on their level.

15:38Altman runs a hand through his short, wavy hair. No disrespect, Elon, but you're not helping us succeed. You poached one of our best researchers to work for you at Tesla. That's a symptom of the problem with OpenAI, not the cause. If OpenAI was doing groundbreaking research on par with Google, I wouldn't have been able to persuade him to come to Tesla. Brockman shifts in his seat with a sigh. Okay, so what's your solution, Elon? What do you think we should do? I think I should take over the company and run it myself. I have the experience and the acumen to turn this company around. Brockman raises his eyebrows skeptically.

16:17You don't think that you already have your hands full running Tesla and SpaceX? No, I believe I can do this and it's the best path forward. Altman stands up. Well, thanks for stopping by, Elon. We'll take your offer under consideration. As soon as Musk is out the door, Altman turns to Brockman. We're not turning OpenAI over to Elon. We absolutely are not. When OpenAI turns down Musk's offer to run the non-profit, Musk leaves the board and fails to honor the rest of the$1 billion he'd pledged, leaving OpenAI in dire financial trouble.

17:03It's spring 2018 in San Francisco. Altman paces his office in OpenAI headquarters while Brockman sits nearby. Altman reaches one end of the office and pivots back. I know we've talked about it before, but I just don't see another way forward. And I still say it's risky. Our non-profit status is one of the key elements that signaled our commitment to ethical AI. It made it possible for us to get those top researchers. Altman sits down in the chair next to Brockman. Look, Elon wasn't wrong that we're in danger of falling behind Google. They've made a huge breakthrough with their transformer neural network.

17:41It trains faster and it has higher accuracy than any previous neural network. Altman and Brockman know that making that kind of leap requires an enormous data set to use for training and a massive amount of capital to pay for even more computer processing power. Altman turns to Brockman. We can't deny it any longer. We need capital. And what's the fastest way to raise a lot of capital, Greg? Bring in investors. Yeah, we can't do that as a non-profit. I know, I know, but our non-profit status sets us apart, right? If we abandon that, are we any different than Google and Facebook, or are we just smaller and poorer?

18:22I don't think it's so black and white. There are ways we can shift to a for-profit entity and still hold on to our ideals. For example, I won't take any equity in the company. I make plenty of money as it is. So that way I won't be making decisions about AI that are financially motivated. And obviously, we'll still be very conscientious about what we release. Our focus on AI that's good for humanity, that's not going to change. We're definitely walking a tightrope here. I ain't saying we're a business, but also that we're not really interested in making money. But I guess if our only other choice is to shut down OpenAI, I really believe that's the choice.

19:03All right, all right. Well, let's discuss it with the other founders and see what they think.

19:10Ultimately, OpenAI abandons its non-profit status and restructures as a for-profit business. It's a big, bold gamble on whether they'll draw investors or repel them. The question is, have the OpenAI founders removed themselves from the race entirely? Or given themselves a shot of steroids?

19:52This holiday season with Manage Engine, their cybersecurity solutions ensure that your data is secure, even while your office lights are turned off. With 18 data centers around the world and over 20 years of experience securing global enterprises, Manage Engine can keep your business running safely, no matter how busy or hectic the world gets. This holiday, gift yourself and your business peace of mind. Visit cybersecurity.manageengine.com. That's cybersecurity.manageengine.com.

20:48Apply for Apple Card in the Wallet app to see your credit limit offer in minutes. Subject to credit approval, Apple Card issued by Goldman Sachs Bank USA, Salt Lake City branch. Terms and more at AppleCard.com.

21:13It's May 2018. team. Google CEO Sundar Pichai strides onto the stage at the Shoreline Amphitheater in Mountain View, California. 7 ,000 people fill the seats cheering for Pichai as if he's one of the rock stars who regularly play this venue. Because in the tech world, Pichai is one. He joined Google in 2004, shortly after receiving his graduate degree from the University of Pennsylvania's prestigious Horton School of Business. He worked on almost all of Google's major products, including Gmail, Chrome, Drive, and Android. And three years ago, in 2015, Pichai was named CEO. Today, he's here to deliver the keynote address for Google's annual I.O.

21:59conference, where he'll announce all the new products and services coming to Google. Pichai waves to the crowd, wearing an olive green fleece and crisp blue jeans. Good morning. Good morning. When the applause finally dies down, Pichai starts off with a joke about emojis. Then he begins methodically going through the product updates to each of Google's services. Pichai talks about how they're redesigning Gmail, maps, photos, and Android to incorporate AI. People clap dutifully, but nothing he's announcing is all that exciting. Until about a half hour into his presentation. We want to connect users to businesses in a good way.

22:47Businesses actually rely a lot on this, but even in the U.S., 60 % of small businesses don't have an online booking system set up. We think AI can help with this problem. Let's say you want to ask Google to make you a haircut appointment on Tuesday between 10 and noon. What happens is the Google Assistant makes the call seamlessly in the background for you. So what you're going to hear is the Google Assistant actually calling a real salon to schedule the appointment for you. Let's listen.

23:24How's something out for you? Hi, I'm calling to book a women's haircut for a client. I'm looking for something on May 3rd. Sure, give me one second. Mm-hmm. Sure. What time are you looking for, Ralph? At 12 p.m. We do not have a 12 p.m. available. The call continues. Google Assistant negotiates a time-in-the-window request, confirms the service the client is looking for, and gives the customer's name, easily responding to questions from the human receptionist at the salon. Okay, perfect. So I will see Lisa at 10 o 'clock on May 3rd. Okay, great. Thanks. Great. Have a great day. Bye. As the recording of the call wraps up, the crowd cheers wildly.

24:09Autocorrect, instantaneous translation, and photo tagging are all uses of AI that have made life easier. But a robot calling and making an appointment on your behalf feels like the AI that was promised in the movies. The kind where robots replace humans. Pichai grins proudly. He continues on with the presentation, but his point has been made. Google's investment in AI is paying off. They're the company that's letting people live in the future. And he makes it seem helpful and benign, a far cry from killer robots and automated drones. But while Google is taking a victory lap, Facebook is staring down yet another setback.

24:56This time, a tragic one.

25:03It's March 2019 in Menlo Park, California. Facebook's chief technology officer, Mike Shrepfer, forces himself to watch a video on his computer monitor. He watches as the camera enters a mosque and opens fire on the worshippers within. Tears stream down his face. This video is less than 24 hours old, and Shrepfer has already watched it several times. He has to, to make sure a video like this never makes it onto Facebook again. A day earlier, a 28-year-old white supremacist walked into two mosques in Christchurch, New Zealand, and shot more than 90 people, killing 51 of them. He had broadcast the shootings using Facebook Live.

25:50The company's AI failed to flag the video as a depiction of graphic violence, and it stayed up on the site for an hour. It's a huge black mark for Facebook. For the past year, Mark Zuckerberg had been touting AI's capabilities in removing banned content before it was posted. And in one of the most crucial moments, it failed. The video ends. Shrepfer wipes a tear from his cheek. Come in. The door opens and one of the engineers on Shrepfer's team walks in. His face looks haggard and taut. I just finished watching the video a bunch of times, and I have some thoughts about why the AI missed it. Thanks for doing that.