Quantifying Urban Revitalization

MIT News, 10/24/2016

Researchers have developed a system that estimates how safe visitors will perceive an area to be based on photographs of the area. The researchers began with a crowdsourcing effort to build an initial database of images and safety ratings of the area in the image. Over 1.4 million responses have then been used to train a machine-learning algorithm to identify these aspects automatically.

Large DDoS Attacks Cause Outages at Twitter, Spotify, and Other Sites

TechCrunch, 10/21/2016

A recent large scale DDoS (distributed denial of service) attack on DNS provider Dyn caused many sites to become temporarily inaccessible, including Twitter, Spotify, and a number of other important social media, e-commerce, news, and code-sharing sites. The attack was coordinated using a large group of compromised devices known as a “botnet”. Rather than computers, much of this botnet consisted of infected IoT (Internet of Things) devices such as security cameras.

How Virtual Reality is Being Used to Deliver Mental Health Care

Australian Broadcasting Corporation, 10/11/2016

Through the use of VR (Virtual Reality), scientists seek to lower the cost and increase the availability of mental health care. They use VR technologies to help patients overcome their fears in a safe environment, by creating simulated situations that reproduce their fears. The availability of this treatment is increasing due to the lower costs of VR devices. The success of the VR experiments might potentially lead to higher quality and more accessible mental health care.

Big Data Help CIA Predict Social Unrest 5 Days Before It Begins

Tech Times, 10/7/2016

The CIA and their new Directorate of Digital Innovation are working on “anticipatory intelligence” to predict future events. The Deputy Director says that they can sometimes forecast outbreaks of unrest up to five days ahead. These predictions are made by using classified information as well as open source data; commentators speculate that much of the data comes from massive social media surveillance.

Google Weakens Allo Privacy Promises

Naked Security, 9/21/2016

Google’s new messaging app, Allo, has been criticized because the default settings provide substantially less privacy than Google had previously announced they would. In part, this is to provide more training data for a “smart reply” feature that generates suggested responses. When messages are stored, law enforcement agencies will be able to access them with warrants.

Was New York’s Mass-Text Manhunt Really Unprecedented?

The Verge, 9/20/2016

New York City police used the Wireless Emergency Alert system recently to send out a “Wanted” text message about a bombing suspect, and they plan to use the system for similar purposes in future. However, they have received heavy criticism, mainly saying that the short, pictureless message may have encouraged mass racial profiling, and that overuse of the system might lead to people ignoring it.

A Beauty Contest Was Judged by AI and the Robots Didn’t Like Dark Skin

The Guardian, 9/8/2016

Beauty.AI developed a set of algorithms to judge photos according to five factors in human standards of beauty; it disproportionately chose photos of white people. The article discusses the potential consequences of emergent bias in algorithms and/or datasets in general, including more consequential examples like predictive policing.

The Ad-Blocking Browser That Pays the Sites You Visit

Wired, 9/1/2016

The Brave web browser (released earlier this year) allows only ads that don’t track users from site to site; it has now added a feature to record how much time users spend on different sites and allow them to send micropayments to those publishers. The article raises examples of how the web has impacted the economics of media and publishing, and also touches on online tracking, data anonymization and de-anonymization, and even Bitcoin.