 So we spend so much time designing efficiency into our products and services, and this is something that we've become really good at. So I want to tell you a little bit about a story that I've been tracking over the last year to show you that we should also be designing for trust. Now Free Lunch is a non-profit project based in China, and they take donations from the public to feed school children in rural parts of China. Now this idea originated with a journalist named Zheng Fei who made a name for himself exposing stories about child kidnapping and organ harvesting. So back in early 2011 he met a teacher from a really rural part of China who told him that she was too embarrassed to eat her lunch in front of her students because they were always hungry. So Zheng Fei was pretty shocked to find out that students were going all day in school just on the breakfast that he ate in the morning before they walked the one to two hours before school. So he went and investigated and he went to several local schools and he found out this wasn't just a local problem, it was country-wide. Millions of children in China were starving all day throughout school. So this problem is also compounded by the demands for China's economic growth. Now a lot of parents are working in the city as migrants, so this means that a lot of children are left alone with little oversight to take care of themselves. So compounding the problem of malnutrition. But Zheng Fei grew up in a village, and this is, here we are in his old home, and he knew the hardships, what it's like to live in a rural life. But he also knew that for children to succeed, they have to be fed. Now as a journalist, he knew about this micro-blogging platform called Sino-Weibo, that much like Twitter allows you to compose messages with 140 characters. But keep in mind, with the way the Chinese language is structured, you can say a lot in 140 characters. So Zheng Fei had witnessed the increasing role of Sino-Weibo as an alternative source for information outside of government-controlled sources. Now more importantly, you notice that people weren't just passively consuming information, but actively sharing censored media, contesting official reports, and demanding feedback from companies. So for many citizens, Sino-Weibo is one of the first places where they can share their opinions publicly without fear. So he had a vision of harnessing this outlet by creating a massive crowdsourced online and reporting donation system that would be based around the idea of trust. Because well aware of the need to make this program free lunch trustworthy, Zheng Fei implemented a combination of online and offline tactics that would make it difficult for schools to embezzle any funds and built around the idea of making everything transparent. So first he had this plan. Participating schools would be given funds to build a kitchen, hire a cook, and buy fresh food. But, in exchange, they would all have to start their own individual Sino-Weibo account and then post a daily report of their financial budget, so every single day. And this way anyone who wanted to check on the financial doings of the school, they could just follow their account directly. And second, each school would assemble a group of locally retired village officials like this 57-year-old farmer to monitor the school and who would have the power just to show up anytime unannounced to look over the school's budget or to, you know, observe the children. So with this plan in place, Zheng Fei sent out his first post with pictures from children from the rural countryside asking people to donate three yuan. So, and three yuan gets, you know, one child, one lunch. So as he reposted, and more people started reposting, forwarding around, and as more schools started sending pictures and other financial reports, that donor, the people who started donating just grew. I mean, it was just an exponential growth. So his plan worked. But here's the crazy part, and this is where it gets kind of interesting, which is that these schools are really hard to get to. I mean, it took us like two to three days just to get to this one school. So there's like no internet and no computers at the school. And even if there were computers, the risk of it being stolen are so high. And here's this one teacher who told me that she is so afraid of her stuff getting stolen that she carries all the kitchen utensils on her back every day, bowls, chopsticks, steam, rice cooker, steamer, everything, one hour between her home and the school. And while all teachers have a cell phone, because as we know there's a mobile revolution in China and all of them text, but they don't have internet. So none of them know how to use the mobile web and much less even go to Sina Webil. By the time it went to the program, none of them had even checked out the Sina Webil website. So these were the conditions that Deng Fei was working with. So here's what he did. And I'm going to walk you through just a quick one day of what it takes to make one single post. Okay, so first, the cook purchased the food and the cook gets the receipts. And then the cook returns to school and then reweighs all the food in front of a witness, like the monitor. And then the monitor and the witness then record the cost and the weight of food onto a piece of paper that is hung up on the kitchen wall. And then after the cook makes the food, the teacher takes the cell phone and takes a picture of the food that the teacher has as a camera phone. And then when the children start getting in line for lunch, a village monitor shows up to kind of make sure that the food, the cook isn't, you know, embezzling or taking any of the food and treating all the kids equally because, you know, you don't want to make sure you're treating the relative of your relative's child with more food. And then while the children are eating, because finally they get their food after this whole long day process, the teacher then gets the accounting book and then takes the paper from the kitchen and then records with a pen all the costs into the accounting book. And then she takes a half hour to one hour to write the 140-character scene awayable post on a text message. But that's if the phone doesn't shut down because, you know, these are Sean's iPhones and if it takes more than a half hour because that's how much time the kids have to eat lunch, then she would have to save the draft and write it after the school days over and that's if the draft isn't deleted because these, you know, phones kind of sometimes shut down. But then she sends it to another monitor, okay, who then has to review the message and that's if the monitor isn't in the field like working or, you know, there's a stable cell phone message so that's assuming that the monitor gets the message and there's nothing to correct, then the monitor sends it to someone in a city, this person called the Urban Volunteer because the Urban Volunteer is an urban area where there's internet access. And if everything looks okay, then the Urban Volunteer posts the financial report to the school's individual scene awayable account. So here's what a post looks like. There's, you know, includes information about the food, the weight of the food, the cost of the food, the budget balance and of course a picture of the food to show proof that was eaten. So this is a whole, this process is so complex and there's so many contingencies and sometimes a single post can take a whole day. But the schools are willing to go through this tedious plan because they risk losing any, all their funding because there's any financial inconsistencies. Now if a designer or any kind of data scientist so like most of us in the room were to look at this, there would be plenty to criticize, okay? Because there would be like, okay, there's plenty, the data's not normalized, the data integrity is compromised and there's a lack of clear and consistent workflows, there's unresolvable redundancies, there's poor search and filtering and there's a lack of naming standards anyway, the list goes on. This works, it really works. Over one million citizens have donated over 30 million U.N. to the program. That's 3.7 million euros and about five million dollars. Over 400 schools are participating in the program, over 14 provinces and even pride of it enterprise have gotten involved. Now donors can donate through TABA which is the largest online shopping site that's equivalent to eBay and over 120 companies have started donating and even the government has gotten involved, they have implemented a similarly structured program like FreeLunch investing 16 billion. So it's not just individuals who are donating, it's government and also private enterprise which speaks to the reach of FreeLunch. Now one of the reasons why FreeLunch works so well is because they really encourage donors to question the school's budgetary spending. So in this example, when a school requested funds for a potato cutting machine, one donor said, from today on if I donate one more dime to FreeLunch I curse myself to hunger for the rest of my life. I saw the budget line item for a disinfection machine and the vegetable peeler, fuck, is that the way you spend the money? I donate and now you pursue modernization instead of just supplying food, you hire five chefs for a kitchen, what are they doing, they can't fucking peel, the students are already eating four dishes a meal, college students don't even have such high standards. So, and by the way, this is 140 characters in Chinese, okay? So then Dengfei explained to the person, okay, so we have 500 students and so we need a potato cutting machine in order to make sure the lunch is served in time. So Dengfei gets like hundreds of messages like these but he sees it as his responsibility to reply to each one and it's an opportunity to also build trust with the donors. So the point is that FreeLunch's massive self-reporting system had already been designed to account for the largest risk possible, which is people's distrust. Now I became really fascinated by the way trust is manifested and the way we use technology. So I started looking back into the history of designing for trust and I found the precedents were in of all places. The standardization of the measurements of electricity. So I want to tell you a little bit about the story. So when electricity first became commercially viable in the late 19th century, scientists knew how to harness it. They knew what to do with it but they didn't know how to measure it. Do you measure the actual energy consumed or do you measure the weight of electricity or do you measure like the light coming out of these bulbs? Or do we use the British Associates of Sciences BA unit otherwise known as the O-Unit or do we use Germany's Siemens unit? So these were the questions that they were asking back then. This very simple thing that we rely on that powers all of our devices was actually really unknown and mysterious and it was just as controversial and debatable as the measure of a social media mention. And there were extreme pressures to figure out. I mean, this was the age of measurement. Every scientist was staying up late in their laboratories at night trying to measure steam power and telegraph communication and perhaps hardest of all electricity. And one of these electricity measuring scientists was the Irish-worn English mathematical engineer William Thompson First Baron Kelvin otherwise known as our Lord Kelvin. Now he upped the pressure on measurement when Lord Kelvin made this statement during a lecture in 1883 is when you cannot measure it when you cannot express it in numbers your knowledge is of meager and unsatisfactory kind. I'm sure that sounds much more scary if I were to say it with a British accent. So upon hearing the statement popularly known as the Curse of Kelvin scientists started freaking out. They felt compelled, even obligated to start measuring the things that they were doing in their labs essentially what the Curse of Kelvin says is that you know, if you're a scientist and whatever you're doing in your lab if you can't measure it you're just some, you know, tinkering around in your lab. So electricity, you know, engineers for example they just got straight to it. They started creating all these devices to measure electricity. So we had, you know, the second meter, the ammeter, the o-meter, the votes meter, the watt meter and more. So what they were all trying to do was they were trying to create accurate devices to measure electricity but most of them never made it to the market and of those that did they eventually failed. So here are some examples. We have the Thomas Miro-Galvin meter the Electrodynamometer and Cisseco meter. Now these devices were all very accurate but they required a measuring specialist to come to your house to take the measurements and interpret the meter through deduction or comparison methods. So this meant that if the consumer didn't trust or didn't know the measuring specialist personally they weren't going to trust a stranger to then measure and then charge for the electricity and to the consumer they couldn't even see what was going on in these devices because to the end user this was opaque. It was a black box. So these devices were made to be very accurate but they were not designed to be transparent and therefore they did not engender trust. So it was at this point that a whole new market of direct reading electricity devices became available. So here we have the Edison meter. So notice that its doors can open allowing the consumer to feel as they could see what was going on in the meters. It looks confusing. Here's the Johnson and Phillip direct dial meter. The dial indicates how much energy was consumed. Now both were designed to make the end user feel as if they could monitor their electricity consumption. They could participate in the process. But here's the thing. These devices were actually way less accurate than the previous devices on the market. But here's the thing that made them successful. They were actually more transparent and because they were more transparent they were more user friendly for the consumer and because of that they were able to engender trust. So ultimately the direct reading dial won because it was the most transparent. So the design innovation of the direct reading dial was born out of a compromise between suppliers, consumers and engineers who had little trust for each other though this distrust created the constraints in which companies designed more user friendly devices. So let's go back to free lunch and we can imagine that if Deng Fei had gone a bunch of internet entrepreneurs together they could have made their own app with the API and then with the smartphone app and it would have all been way more efficient and way more accurate than that crazy process that I had showed you but to the end user in China it would have felt like a black box. So the program's overlapping layers of redundancy creates a system of checks and balances that does two things. It may corruption more difficult while it valorize a culture of data keeping and monitoring. So with schools posting their daily financial reports to Sina Weibull the measures that were not some centralized group of with authority, the measures became everyone. Anyone with access to Sina Weibull can now monitor their accounts. So this is a really radical shift in China because it means that the direct measure is not coming from the government but from the people. So we're seeing a collective social shift in the way information feedback and accountability conceptualize as something of value in China. So by building free lunch around the idea of trust in a society that sees this as something new this is a really big deal. So for designing for trust and this is not just for users in China but users everywhere and especially for those of us who are designing more participatory and more social media oriented platforms because communities on these platforms would collapse without trust or would be unable to form. There's some design principles that we should be looking at which is first, lower the threshold for your users to establish trust. Make it easy for them to judge the veracity of information sources. Conduct a thorough ethnographic study on how users conceive of information and trust because trust practices will vary in different cultures and that will affect your actual end product. And you don't have to be an ethnographer to do this. Airbnb's Brian Chesky, the co-founder he lives in his service and he's also an informal ethnographer so you can start doing this today. Ask what user-centered values you want to prioritize such as transparency or familiarity and treat those values as healthy constraints for innovation, not against it. But at the same time, be clear about what values are being compromised because understand that some design compromises are only appropriate for some contexts. You know, when a community may be okay with a compromise in efficiency or accuracy but it may not be okay for another. And design minimally enough so that you can watch what user-centered values emerge out of the interaction. And lastly, avoid the curse of Kelvin because just because something isn't quantifiable doesn't mean that it doesn't have value. And these principles become even more critical for teams of designers, social scientists, and technologists working in cultural contexts that they may not understand as well. Now, I say all this having just returned from two years of field work in China where I traveled over 60,000 miles on trains and planes and boats and scooters and donkeys and all to figure out, you know, to emerge how people use technology on the ground. And I used really ethnographic, immersive ethnographic methods like living with gamers inside internet cafes, working with migrants as a street vendor so that I could observe how secondary cell phone markets pop up around construction sites or shopping with teens so that I can understand emerging consumer habits. So in the last year, we've seen some really eye-opening economic statistics about China that Geraldine also mentioned. You know, it's already the world's second-largest economy. We have the largest internet users and social media users. But what these statistics don't tell you is that millions of people in China are acquiring, forwarding, and sharing information with people they don't personally know and some of this information is actually quite sensitive. So it signals to a massive and collective shift of people's trust moving from more close to open networks. And this opening creates opportunities for programs like Freelunch and many others to exist. Now with all, while these cultural and social conditions may be very unique and specific to China, the global trend is that our platforms and digital services are becoming more crowd-based and participatory. So that means there's going to be a whole lot of data sharing, creation, and processing happening. So we may find ourselves asking the same questions as the electricity scientists and engineers of the late 19th century, which is how are we going to harness all this data? Or now that we know how to harness it, what are we going to do with it? And now we know what to do with it. And here's a big question. We have a lot of it now. How are we going to measure it? So in a world of big data, it's going to be continually seductive to think that we have finally figured out a way to measure human emotions. I mean just recently yesterday, conference co-speaker Lauren Anderson from collaborative consumption told me about a trust cloud where you can quantify your online reputation to a trust score, like a credit score, or similarly legit, which creates a reputation score that follows you across sites. So right now, we have our own curse of Kelvin in thinking that data is more valuable once it is calculated and normalized. And essentially meaning can be monetized. But this is a self-imposed curse. And we can lift this curse if we're willing to embrace a new design paradigm. And this paradigm may have criteria that's not just technical but social. Historically, even technical efforts have seemed incredibly quantifiable like the measure of electricity weren't solely built on the principles of science, but emotions. And our interactions are driven much more by our emotions than we like to admit. But look, this is how trust works. So for designing for trust between humans and computers, it requires us to prioritize users' needs not just the system's needs. Thank you.