The city of Austin wants to find out. It’s piloting a new blockchain platform to improve identity services for homeless citizens and ensure they possess their own information. Identity is a tough problem to tackle, and it gets harder with marginal populations like the homeless or refugees. A loss of personal ID possessions typically means having to start from scratch, making it harder to transition out of homelessness.
There are about 2,000 homeless people in Austin and several thousand at different stages of transition, so city officials want to integrate its government and private homeless service providers to improve services. This means verifying the identity of a person looking for help, knowing previously received care, and allowing people to own their own records. The goal is to consolidate the identity and vital records of each person safely, while allowing service providers to access the information.
Using blockchain can also replace paper records with electronic encrypted records and create a decentralized authentication mechanism to verify identity. Because records on blockchain could build over time, different providers would know an individual's service use history — which would be particularly helpful in health care. TechCrunch
Did You Hear? Mark Zuckerberg Testified Before Congress
Yes, he was asked many questions about how Facebook works and responded to lawmakers’ complaints about the company’s privacy practices. But he also talked about artificial intelligence and how it can help solve many of Facebook’s problems. In fact, he brought up AI more than 30 times during the 10 hours of questioning last week. One day, he said, AI will be smart enough to fight against fake news, hate speech, discriminatory ads and terrorist propaganda.
Zuckerberg said this could happen in the next five to 10 years, adding that AI can even help with censorship, fairness and human moderation. When he created Facebook in his dorm room, the AI technology to identify online platform misbehavior didn't exist, but looking ahead, building AI tools will be the scalable way to identify harmful content.
And while Facebook does have AI technology, it can’t do any of this well enough yet. It’s not clear when it’ll be able to. Some tech experts think Zuckerberg wants to use AI to help the company “pawn off blame from the humans creating it.” Still, Zuckerberg is optimistic that in five to 10 years, Facebook’s AI will understand the “linguistic nuances” of content well enough to catch potential risks. The Washington Post
NASA Needs Help from SpaceX
The space agency’s latest exoplanet-hunting spacecraft, the Transiting Exoplanet Survey Satellite, or TESS, is rescheduled to launch April 18. It has a 2-year mission to search the sky for habitable worlds, and is getting a boost from SpaceX to do so. TESS is launching atop of the SpaceX Falcon 9 rocket, which successfully test-fired April 11. The Falcon 9 will use a new first stage reusable booster for this flight, and once it does its job, it’ll return to Earth on a drone ship in the Atlantic Ocean.
TESS will survey the sky for small “dips” in the light coming from nearby stars, which indicate a planet could be passing in front of its parent start. Astronomers call this a transit method, which the Kepler Space Telescope also used to identify thousands of candidate planets outside our solar system.
But the mission scientists expect more from TESS, as it will survey a patch of sky hundreds of times larger than Kepler did. And it needs the Falcon 9 to help it reach a particular orbit by intercepting the moon and using lunar gravity to get there. But the launch is tricky; the moon and TESS need to be in the right place so the lunar-gravity assist can happen. The original launch window was 30 seconds long, so fingers crossed! Wired
An Exhibit of Citizens’ Secrets
Chinese artist Deng Yufeng bought the personal information of more than 300,000 Chinese people from the internet and displayed it in a public exhibition in the city of Wuhan called “346,000 Wuhan Citizens’ Secrets.” Why? He wanted to prove a point: you don’t have the data privacy you think you do. But it didn’t go over so well to police. After two days, local authorities shut down the exhibit and told Yufeng he would be investigated on suspicion of obtaining this information illegally.
There’s a growing debate about the lack of data privacy in China. Online brokers are buying and selling information online, and Chinese people are receiving constant unwanted calls and text messages. In fact, Yufeng used the Chinese messaging app QQ to find sellers of information, and realized how easy the data was to find.
He paid a total of $800 for people’s names, genders, phone numbers, online shopping records, travel inquiries and license plate numbers. He printed the information on sheets of paper, redacting key identifying details, and hung them on a museum wall. Though he’s under investigation and could face up to seven years in jail, Yufeng’s artwork did gain media attention, and represented the people’s push back against tech companies. The New York Times
Virtual Robots are Teaching Themselves Kung Fu
AI researchers at UC Berkeley and the University of British Columbia created virtual characters that copy the way a person does martial arts, parkour and acrobatics. They even practice the moves until they get them right, so this could transform the way video games and movies are made. Rather than planning a character’s actions, animators might soon be able to input real footage into a program and have the characters learn, practice and master the moves on their own.
They use the AI technique called reinforcement learning, which is modeled after the way animals learn. The researchers used the actions of expert martial artists and let the virtual character experiment with the moves, providing positive reinforcements with improvement. The same algorithms can be used to train a character to perform other human-like moves too, like backflips or dances, rather than programming computer-generated characters with time-consuming choreographed scripts.
This can be a game changer for the animation and computer game industries as well, as both are already exploring the use of software that adds realistic physics to characters. In time, even real robots could learn to perform complex tasks with simulated practice, or teach humans new tricks. MIT Technology Review