How Singapore lost trust on contact tracing privacy

Jumbled up sentences show that AIs still don't really understand language 

Sponsored by Siemens Digital Industries Software The Download Your daily dose of what's up in emerging technology Broken promises 01.12.21 Good morning! Today: how Singapore misled its citizens over contact tracing app data, new research demonstrates AI still doesn't understand what it's talking about, and inside China's high-tech rural agriculture sector. Get your friends to sign up here to get The Download every day. How Singapore lost trust on contact tracing privacy   Promises, promises: For Singaporeans, the covid-19 pandemic has been closely intertwined with two technologies: the QR code, used to check into public venues, and the country’s contact tracing app, TraceTogether. When they launched, there wasn’t much space for the public to discuss their concerns. But the government repeatedly told Singaporeans that the data collected with such technology would be used only for contact tracing during the pandemic.  Ah, but: Earlier this month, it emerged that the government’s claim was false. The Ministry of Home Affairs confirmed that data could actually be accessed by the police for criminal investigations. These revelations triggered public anger and criticism, not necessarily because Singaporeans are particularly privacy conscious—in fact, state surveillance is largely normalized in the country—but because people felt they’d been subjected to a bait-and-switch. The response: The government has since announced that it will introduce new legislation to limit law enforcement’s use of contact tracing data. But there is no timeline yet as to when the proposed legislation will be brought before parliament, and details are scant. Read the full story. —Kirsten Han Jumbled up sentences show that AIs still don’t really understand language The news: AIs don’t really understand language at all. Researchers at Auburn University and Adobe Research tried to get an natural language processing system to generate explanations for its behavior, such as why it claimed different sentences meant the same thing. When they tested their approach they realised that shuffling words in a sentence made no difference to the explanations. The details: For example, the systems correctly spotted that the sentences “Does marijuana cause cancer?” and “How can smoking marijuana give you lung cancer?” were paraphrases. But they were even more certain that “You smoking cancer how marijuana lung can give?” and “Lung can give marijuana smoking how you cancer?” meant the same thing too. Why? The models appear to pick up on a few key words in a sentence, whatever order they come in. The good news is that it might not be too hard to fix. The researchers found that forcing a model to focus on word order, by training it to do a task where word order mattered, such as spotting grammatical errors, also made the model perform better on other tasks. Read the full story. —Will Douglas Heaven China’s rural agriculture doesn’t just feed people—it powers the future Blockchain Chicken Farm, a new book from Oakland-based writer, designer, and scholar Xiaowei Wang, explores technology in rural China and the surprising ripple effects of the country’s food supply chain on people all around the world. The book connects, for example, an AI-driven pig-farming operation in Guangdong to Silicon Valley surveillance culture. It also includes a selection of speculative “Sinofuturist recipes,” an ongoing art project that uses food to address anxieties about technology, the ecosystem, and the body. We discussed Wang’s research, the effects of the coronavirus pandemic, and what China’s food system means for us all. Read the full interview. Read next: a cocoa producer in Mexico explains how automation has transformed the process. And: how sexual reproduction might help us to find new potato varieties for the first time in decades. These three articles are from the latest issue of MIT Technology Review, all about food. Read the full magazine, and if you haven’t already, you can subscribe to MIT Technology Review for the year for as little as $50. We can still have nice things A place for comfort, fun and distraction in these weird times. (Got any ideas? Drop me a line or tweet 'em at me.)   + UNLIMITED PUPPIES (1,000 thank yous to Stefan for this gift). And just look at this little cutie.    + This Chinese restaurant in Montreal is serving up TRUTH as well as (allegedly delicious) food.   + A baby wombat enjoying some flowers. Awww.    + Might have to change my name.    + Music to make you love the flute. (NYT $)   + Struggling to cook rice on the stove? Try popping it in the oven.    + Jealous of all those companies getting to ban Trump? This game lets you get in on the action. Sponsor Message With trust in AI, manufacturers can build better Some people might not associate the word “trust” with artificial intelligence (AI).   Stefan Jockusch, Vice President of Strategy at Siemens Digital Industries Software, is not one of them. AI can improve production speed, take away tedious tasks from creative engineers, and make products customers want, explains Jockusch. He gives the example of Apple’s iPhones and its facial recognition software—technology that has been tested “millions and millions of times” and produced just a few failures.    Listen to the podcast here. The top ten must-reads I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology. 1 What's going wrong with the US's vaccine program A fractured approach will not get us out of this mess. (WP $)   + See how it's going in your state. (NYT $)   + Researchers are debating the merits of different dosing strategies. (Nature)   + What it will take for the UK to vaccinate 14 million people by mid-February. (Wired UK)   + We're not going to reach herd immunity in 2021, says the WHO. (Deutsche Welle)   + BioNTech is confident its vaccine will work against new variants of covid-19. (CNBC)  2 Every single Parler post has been archived  Mostly thanks to one researcher's tireless efforts. (Gizmodo)   + The scramble to archive Capitol insurrection footage before it disappears. (TR)   + Parler is suing Amazon for cutting it off. (WSJ $)   + On Telegram, white nationalists are trying to radicalize ex-Parler users. (Mother Jones)   + Amazon has started removing QAnon merch. (WP $)  3 Angela Merkel does not like the fact Twitter has suspended Trump Seems politicians might be finally waking up to the merits of reining in Big Tech. (CNBC)   + Nor does Russia's opposition leader. (NBC)   + Why world leaders are concerned. (Quartz)   + One person showed up to protest against the decision at Twitter's HQ. (The Verge)   + Twitter has purged over 70,000 QAnon accounts. (WP $)   + An internet provider was worried about censorship… so it blocked access to Twitter and Facebook and required users to opt in. (Newsweek)  4 Facebook has shut down Ugandan government officials' accounts Seems it's getting a taste for banning dodgy politicians. (Reuters)   + Facebook's oversight board should review Trump's suspension. (Lawfare)   + It says it will crack down on "stop the steal" content. (The Guardian)  5 Telegram and Signal downloads are surging WhatsApp has updated its terms to make it clearer how it will share data with its parent company, Facebook. (CNBC)   + Why Signal is the best bet for privacy. (The Next Web)  6 Inside China's toxic tech workplaces A woman's death has fuelled an employee backlash against unreasonably long hours and public humiliation. (Vice)   7 To survive, plan for the worst Companies should do this too. (Wired $)   8 How robots might be able to learn empathy (sort of) 🤖❤️ Giving robots the ability to predict others' actions would be another step forward. (CNET)   9 And… here comes that bitcoin price drop 📉 About $185 billion in “value” wiped out virtually overnight. (Bloomberg)   + Pity the millionaires that can't remember the passwords for their crypto wallets. (NYT $)  10 Sea sponges have incredible skeletons The lattice structure has a remarkable ability to withstand compression and bending. (Quanta) We want to hear from you: the global AI agenda 2021.  In just 10-minutes, you can help MIT Technology Review understand how enterprises are using big data and AI to recover from covid-19. This survey also explores how businesses are reevaluating data privacy and citizen consent to AI practices. Your response is anonymous and would be greatly appreciated. You can download last year's Global AI Agenda here for free. Take the survey now. Quote of the day “This is not about finding China guilty." —Fabian Leendertz, epidemiologist at Germany’s public health body and part of the WHO team set to visit China this week, tells The Guardian the mission is scientific, not political. Top image credit: SHUTTERSTOCK Please send puppies to hi@technologyreview.com. Follow me on Twitter at @charlottejee. Thanks for reading! —Charlotte Was this newsletter forwarded to you, and you'd like to see more? Sign up for free Intelligent computing Join us virtually | February 10-11, 2021 Want to change how you receive these emails? You can *update your preferences or *unsubscribe from this list. Subscription Information This email was sent to Webmaster220 Bible Study Blog *the blog owner has deleted the hyperlink intentionally Follow us: MIT Technology Review One Main Street Cambridge, MA 02142