Resources

Google has fired dozens of employees between 2018 and 2020 for abusing their access to the company’s tools or data, with some workers potentially facing allegations of accessing Google user or employee data, according to an internal Google document obtained by Motherboard. From a report:
The document provides concrete figures on an often delicate part of a tech giant’s operations: investigations into how company’s own employees leverage their position inside the company to steal, leak, or abuse data they may have access to. Insider abuse is a problem across the tech industry. Motherboard previously uncovered instances at Facebook, Snapchat, and MySpace, with employees in some cases using their access to stalk or otherwise spy on users.

The document says that Google terminated 36 employees in 2020 for security related issues. Eighty-six percent of all security-related allegations against employees included mishandling of confidential information, such as the transfer of internal-only information to outside parties. 10 percent of all allegations in 2020 concerned misuse of systems, which can include accessing user or employee data in violation of Google’s own policies, helping others to access that data, or modifying or deleting user or employee data, according to the document. In 2019, that figure was 13 percent of all security allegations.

“Free apps marketed to people with depression or who want to quit smoking are hemorrhaging user data to third parties like Facebook and Google — but often don’t admit it in their privacy policies, a new study reports…” writes The Verge.

“You don’t have to be a user of Facebook’s or Google’s services for them to have enough breadcrumbs to ID you,” warns Slashdot schwit1. From the article:
By intercepting the data transmissions, they discovered that 92 percent of the 36 apps shared the data with at least one third party — mostly Facebook- and Google-run services that help with marketing, advertising, or data analytics. (Facebook and Google did not immediately respond to requests for comment.) But about half of those apps didn’t disclose that third-party data sharing, for a few different reasons: nine apps didn’t have a privacy policy at all; five apps did but didn’t say the data would be shared this way; and three apps actively said that this kind of data sharing wouldn’t happen. Those last three are the ones that stood out to Steven Chan, a physician at Veterans Affairs Palo Alto Health Care System, who has collaborated with Torous in the past but wasn’t involved in the new study. “They’re basically lying,” he says of the apps.

Part of the problem is the business model for free apps, the study authors write: since insurance might not pay for an app that helps users quit smoking, for example, the only ways for free app developer to stay afloat is to either sell subscriptions or sell data. And if that app is branded as a wellness tool, the developers can skirt laws intended to keep medical information private.

A few apps even shared what The Verge calls “very sensitive information” like self reports about substance use and user names.

In the past couple of years, genetic-testing companies like Ancestry and 23andMe have become popular for finding out family history and DNA information. More than 12 million Americans have sent in their DNA to be analyzed to companies like 23andMe and AncestryDNA. The spit-in-tube DNA you send in is anonymized and used for genetic drug research and both sites have been selling the data to third-party companies, like P&G Beauty and Pepto-Bismol, and universities, like The University of Chicago, for some time. In fact, just last week major pharmaceutical giant, GlaxoSmithKline, announced a $300 million deal with 23andMe. The deal entails that they can use the data to analyze the stored sample, investigate new drugs to develop and genetic data for how patients are selected for clinical trials. Both 23andMe and Ancestry said that they will not share genetic information freely, without a court order, but people are welcome to share the information online themselves sometimes in order to find lost relatives or biological parents.

A French police officer has been charged and arrested last week for selling confidential data on the dark web in exchange for Bitcoin,” reports ZDNet. French authorities caught him after they took down the “Black Hand” dark web marketplace. Sifting through the marketplace data, they found French police documents sold on the site. All the documents had unique identifiers, which they used to track down the French police officer who was selling the data under the name of Haurus.

Besides selling access to official docs, they also found he ran a service to track the location of mobile devices based on a supplied phone number. He advertised the system as a way to track spouses or members of competing criminal gangs. Investigators believe Haurus was using the French police resources designed with the intention to track criminals for this service. He also advertised a service that told buyers if they were tracked by French police and what information officers had on them.

“In the future the data procured from smartwatches might be much more valuable than what is currently available from laptop and mobile users,” reports David Curry, raising the possibility that stores might someday use your past Google searches to alert you when they’re selling a cheaper product.”

Feed the data of millions of people playing various computer games into AI machine learning and shaping algorythms… It’s already happening to an extent:

“The latest computer games can be fantastically realistic. Surprisingly, these lifelike virtual worlds might have some educational value, too—especially for fledgling AI algorithms.

Adrien Gaidon, a computer scientist at Xerox Research Center Europe in Grenoble, France, remembers watching someone play the video game Assassins Creed when he realized that the game’s photo-realistic scenery might offer a useful way to teach AI algorithms about the real world. Gaidon is now testing this idea by developing highly realistic 3-D environments for training algorithms how to recognize particular real-world objects or scenarios.

The idea is important because cutting-edge AI algorithms need to feed on huge quantities of data in order to learn to perform a task. Sometimes, that isn’t a problem. Facebook, for instance, has millions of labeled photographs with which to train the algorithms that automatically tag friends in uploading images (see “Facebook Creates Software that Matches Faces Almost as Well as You Do”). Likewise, Google is capturing huge amounts of data using its self-driving cars, which is then used to refine the algorithms that control those vehicles.

But most companies do not have access to such enormous data sets, or the means to generate such data from scratch.

To fill in those gaps, Gaidon and colleagues used a popular game development engine, called Unity, to generate virtual scenes for training deep-learning algorithms—a very large type of simulated neural network—to recognize objects and situations in real images. Unity is widely used to make 3-D video games, and many common objects are available to developers to use in their creations.

A paper describing the Xerox team’s work will be presented at a computer vision conference later this year. By creating a virtual setting, and letting an algorithm see lots of variations from different angles and with different lighting, it’s possible to teach that algorithm to recognize the same object in real images or video footage. “The nice thing about virtual worlds is you can create any kind of scenario,” Gaidon says.

Gaidon’s group also devised a way to convert a real scene into a virtual one by using a laser scanner to capture a scene in 3-D and then importing that information into the virtual world. The group was able to measure the accuracy of the approach by comparing algorithms trained within virtual environments with ones trained using real images annotated by people. “The benefits of simulation are well known,” he says, “but [we wondered], can we generate virtual reality that can fool an AI?”

The Xerox researchers hope to apply the technique in two situations. First, they plan to use it to find empty parking spots on the street using cameras fitted to buses. Normally doing this would involve collecting lots of video footage, and having someone manually annotate empty spaces. A huge amount of training data can be generated automatically using the virtual environment created by the Xerox team. Second, they are exploring whether it could be used to learn about medical issues using virtual hospitals and patients.

The challenge of learning with less data is well known among computer scientists, and it is inspiring many researchers to explore new approaches, some of which take their inspiration from human learning (see “Can This Man Make AI More Human?”).

“I think this is a very good idea,” says Josh Tenenbaum, a professor of cognitive science and computation at MIT, of the Xerox project. “It’s one that we and many others have been pursuing in different forms.”