DeepTech – Project JEDI & Maven, Creative Potential of AI, Google’s Utopian City Quayside, & more

Creative gains in AI reshapes global labor force

Amid mounting fears of sector automation, there was reassurance that some jobs would be shielded from the inexorable march of the algorithm. Since AI couldn’t replace human imagination, surely creative jobs are safe. Policymakers will now have to consider whether or not AI can learn creativity and if so, what that means for global labor markets. AI-created ‘artwork’ has fetched nearly half a million dollars at auction. Google Assistant has already passed the turing test; now, it is seamlessly multilingual. While AI has written novels and poems, not all the writing is great. AI can write jokes, but they just aren’t funny. (Maybe they’re funny to other AI). MIT researchers’ quest to find out if machines can learn to scare us led them to develop AI-powered horror. MIT’s Nightmare machine and uncanny musicbox both use AI to generate terrifying imagery and sounds. And, MIT’s moral machine attempts to teach AI human moral reasoning, which has immediate, real-world applications for self-driving cars. If MIT’s true aim was to frighten, they didn’t have to develop either nightmare machine or uncanny musicbox. The moral machine is truly terrifying. Think AI can’t be racist, sexist, or otherwise biased? Think again. MIT researchers have also created Norman, an artificial intelligence that sees only death, to underscore the inherent biases in AI development.

• MIT Technology Review: A controversial artwork created by AI has hauled in $435,000 at auction
• Motherboard: Get in the Halloween Spirit with This AI-Generated Spooky Music
• Nature Magazine: Look to Africa to advance artificial intelligence
• MIT Technology Review: Should a self-driving car kill the baby or the grandma? Depends on where you’re from.
• Nature Magazine: Self-driving car dilemmas reveal that moral choices are not universal

Balkanization of the Internet

Recent legislation in Europe, development of Google’s DragonFly, and Russian attempts to build its own internet will most likely balkanize the internet, experts warn. The EU’s General Data Protection Regulation, a massive overhaul of data privacy regulations, went into effect earlier this summer. Google has gone back behind China’s Great Firewall since exiting the country in 2010 over censorship. Google.cn has been blocked since the exit, but Google is working with Chinese officials to develop Dragonfly, a censorable search engine. Last year Russia announced plans to build its own Internet, and earlier this month Russia’s Federal Security Service has called for complete control over cyberspace. In the US, the FCC’s Restoring Internet Freedom Order repealed Network Neutrality and went into effect this June. Over the past decade each country has developed a unique set of regulations on privacy, data localization, and censorship that are quickly becoming incompatible. And, in case you’re wondering whether or not Russia could even create its own Internet, ICANN’s chief technology officer says it’s entirely possible.

• Freedom House: Freedom on the Net 2018: The Rise of Digital Authoritarianism
• Motherboard: A Chinese Dissident Artist Is Taking Aim at Google for Censored Search Engine
• ACLU: What Is Net Neutrality?
• IEEE Spectrum: Why Russia Is Building Its Own Internet
• Fortune: Google CEO Sundar Pichai Finally Opens Up About That Censored Search Service For China
• Defense One: Russian Intel Chief: Internet Should be Controlled By ‘Proper Authorities’
• New York Times: There May Soon Be Three Internets. America’s Won’t Necessarily Be Best
• The Intercept: Google Plans To Launch Censored Search Engine In China
• Forbes: Did Google Bow To China In Refusing To Provide AI Technology
• The Wall Street Journal: The Global Tech Backlash Is Just Beginning

Projects JEDI & Maven: literally the most boring DoD projects ever?

The US Department of Defense will soon have two active tech contracts, Projects JEDI and Maven. The first is like dropbox for the army & the second seeks to do what facebook has been doing for years, analyze photos. In April 2017, the US Dept of Defense launched Project Maven. Maven’s purpose is to integrate AI, big data, and machine learning with the aim of analyzing moving and still imagery (read: analyze aerial drone footage). Part of Maven’s initial work was carried out by Google, who has since declined to renew its contract. After facing backlash for involvement in project Maven, Google has pledged $25M towards the development of ‘good’ AI and has turned down a role in equally controversial Project JEDI. Where Maven was a smaller, narrowly-focused project to analyze drone footage, Project JEDI is a colossal $10B contract  to provide commercial cloud computing for all US Military branches. Among the Pentagon’s stated goals of JEDI, a focal point is to “transform how DoD captures, processes, understands, and harnesses its data to deliver advanced capabilities [and] enable real-time decision-making.” For a combined expenditure of $17B, it’s likely that these projects have significant top secret components. Still think it’s sci-fi? Have a look at the original DoD memo establishing Maven and the Pentagon’s internal strategy document outlining JEDI.

• Harvard Kennedy School Belfer Center: Artificial Intelligence and National Security full text
• Us Dept. of Defense: Establishment of an Algorithmic Warfare Cross-Functional Team (Project Maven) Official Memo
• US Dept. of Defense: Project Maven to Deploy Computer Algorithms to War Zone by Year’s End official press release
• Federal Business opportunities: JEDI Cloud RFP Official Bid
• NextGov: Pentagon’s Next Cloud Contract Could Be Worth Billions
• Gizmodo: Google Plans Not to Renew Its Contract for Project Maven, a Controversial Pentagon Drone AI Imaging Program
• Nature Magazine: Retool AI to forecast and limit wars

Cryptocurrency & Blockchain: 2 completely different approaches

Many of the world’s regulatory agencies are still playing catch up with cryptocurrencies and the blockchain technology underpinning some of them. One reason policymakers find cryptocurrencies problematic is because they lack common features of fiat currencies such as a central bank. This lack of traditional monetary policy has forced many cryptocurrencies to enact their own monetary policies, which are vastly different than the fed raising the interest rate. Ethereum’s planned “difficulty bomb” seeks the same effects of traditional contractionary monetary policy, but instead of changing the reserve requirements or raising the interest rate, a piece of code is added that makes it increasingly difficult to mine blocks (or, units of currency). Japan has officially granted its cryptocurrency industry the power to self-regulate. Japan has established a Virtual Currency Exchange Association, an agency comprised of representatives from cryptocurrencies, and given it the power to levy sanctions. Meanwhile, in China the Cyberspace Administration has proposed restrictions on online firms that use blockchain technology. The new regulations would require companies to gather users’ real names and national identification numbers and turn the data over to law enforcement. Regardless of their stance on cryptocurrencies, both countries recognize the vast potential for blockchain to streamline government and financial services as well as promote transparency. (Once information is stored in a blockchain, it can’t be removed.) Which is the one thing that China isn’t in love with.

• South China Morning Post: China requires blockchain-based information service providers to register users using real names, censor postings and store user data
• MIT Technology Review: Japan just gave its cryptocurrency industry legal power to police itself
• Coindesk: Before the ‘Bomb’ Hits: Why the Race Is on to Alter Ethereum’s Economics
• Intellect Soft: Blockchain Regulation: Technology Is Welcomed, Cryptocurrency Regulated
• Cyberspace Administration of China: Notice of the National Internet Information Office on the Public Consultation on the Regulations on the Management of Blockchain Information Services (Draft for Comment) Official

Welcome to Quayside, Google’s Modern Day Pleasantville

Google’s Sidewalk Labs is developing a city, and insiders are already leaving the project over privacy concerns. Located on a tiny patch of land about a tenth of the size of Vatican City, Quayside is mostly owned by the Canadian state and is expected to have around 5,000 residents. Quayside represents Google’s attempt to create a community with zero-emission, self-driving public transit; underground tunnels for robots to transport mail and rubbish; carbon offsetting electric grids; and expandable, modular structures to ensure affordable housing. Because Google developers see smartphones as excellent analogies to smart cities, there will be sensors everywhere and on every surface imaginable, quietly collecting data. A major issue concerns use and ownership of the data. Some data will be used for tracking the flow of pedestrians and city planning. But, local Toronto residents question exactly “What data will be collected, how personal will it be, how will it be used, and who will have access to it?”  Google plans to allow 3rd party access to developers, and so far their primary business model seems to be licensing the Quayside-developed technology to other cities. The only problem is that cities aren’t smartphones, and the subway isn’t an app.

• Torontoist: Civic Tech: A list of questions we’d like Sidewalk Labs to answer
• MIT Technology Review: A smarter smart city:
• Guardian :‘City of surveillance’: privacy expert quits Toronto’s smart-city project
• Sidewalk Labs: About Sidewalk Toronto